<?xml version="1.0" encoding="utf-8"?>
<?xml-stylesheet type="text/xsl" href="http://feeds.qzone.qq.com/rss.xsl" version="1.0"?>
<rss version="2.0" xmlns:qz="http://qzone.qq.com">
<channel>
<title><![CDATA[wildyouth]]></title>
<description><![CDATA[550115938]]></description>
<link>http://550115938.qzone.qq.com</link>
<lastBuildDate>Wed, 25 Nov 2009 18:27:59 GMT</lastBuildDate>
<generator>Qzone</generator>
<language>zh-cn</language>
<copyright>Copyright (C), 2005-2008, Tencent Tech. Co., Ltd.</copyright>
<pubDate>Wed, 03 Dec 2008 15:39:48 GMT</pubDate>

<item>
<title><![CDATA[中国汽车产业发展]]></title>
<link>http://550115938.qzone.qq.com/blog/1228318788</link>
<description><![CDATA[    <span style="font-size:13px;line-height:1.8em;">中国自称是“<span style="font-weight:bold"><wbr />世界的制造中心</span><wbr />”，我想顶多只能算是“<span style="font-weight:bold"><wbr />世界的加工工厂</span><wbr />”，而且这“世界加工工厂”的梦想也随着金融危机而破灭。但这并不是一件悲哀的事。中国这种来料加工低附加值的制造业其本身抵抗风险的能力是很低的。在宏观调控政策下，收紧银根，提高银行准备金；新的劳动合同法实施，更是增加了企业的劳动成本，最大的低价劳动成本竞争力开始打破；又加上金融海啸的影响，外部需求降低，人民币升级，出口重创。这一切无疑加速了制造业倒闭。这种粗放型的，低技术含量，低附加值的制造业在过去是经济增长的重要力量，但高消耗，低产出，用环境和资源换来的发展，正严重阻碍着中国的进一步发展。<span style="font-weight:bold"><wbr />中国产业必须升级！</span><wbr />我们要有自主知识产权的制造业，不但要<span style="font-weight:bold"><wbr />中国制造</span><wbr />，更要<span style="font-weight:bold"><wbr />中国创造</span><wbr />。</span><wbr /> <br><span style="font-size:13px;line-height:1.8em;">     中国汽车工业的发展是令人悲哀的，国有企业落后的管理和技术我就不想提了，在合资企业中，我们并没有能够消化别人的技术，关键的技术我们更是连消化的机会都没有。所谓的研发也只是抄袭，究其原因是企业舍不得在研发上投入，研发是靠积累的一个过程，需要长期的投入，却不是立刻见效的，需要一个长远的战略眼光，因为这个积累的过程一旦完成，其效益是可想而知的。中国有能力在尖端的航天领域取得瞩目的成就，一定有能力在汽车行业开创自己的未来，这不是能和不能的问题，而是做和不做的问题。</span><wbr /> <!--v:3.2--> ]]></description>
<category><![CDATA[个人日记]]></category>
<author><![CDATA[550115938@qq.com(wildyouth)]]></author>
<comments>http://550115938.qzone.qq.com/blog/1228318788#comment</comments>
<qz:effect>134218240</qz:effect>
<pubDate>Wed, 03 Dec 2008 15:39:48 GMT</pubDate>
<guid>http://550115938.qzone.qq.com/blog/1228318788</guid>
</item>

<item>
<title><![CDATA[MIT solves 100-year-old engineering problem]]></title>
<link>http://550115938.qzone.qq.com/blog/1223141046</link>
<description><![CDATA[MIT solves 100-year-old engineering problem Insights on fluid flow could impact fuel efficiency, moreElizabeth A. Thomson, News Office<br>September 24, 2008<br>As a car accelerates up and down a hill then slows to follow a hairpin turn, the airflow around it cannot keep up and detaches from the vehicle. This aerodynamic separation creates additional drag that slows the car and forces the engine to work harder. The same phenomenon affects airplanes, boats, submarines, and even your golf ball. <br><br>Now, in work that could lead to ways of controlling the effect with potential impacts on fuel efficiency and more, MIT scientists and colleagues have reported new mathematical and experimental work for predicting where that aerodynamic separation will occur. <br>The research solves &quot;a century-old problem in the field of fluid mechanics,&quot; or the study of how fluids -- which for scientists include gases and liquids -- move, said George Haller, a visiting professor in the Department of Mechanical Engineering. Haller's group developed the new theory, while Thomas Peacock, the Atlantic Richfield Career Development Associate Professor in the same department, led the experimental effort.<br>Papers on the experiments and theory are being published in the Sept. 25 issue of the Journal of Fluid Mechanics and in the September issue of Physics of Fluids, respectively.<br>Fluid flows affect everything in our world, from blood flow to geophysical convection. As a result, engineers constantly seek ways of controlling separation in those flows to reduce losses and increase efficiency. One recent accomplishment: the sleek, full-body swimsuits used at the Beijing Olympics.<br>Controlling fluid flows lies at the heart of a wide range of scientific problems, including improving the performance of vehicles, Peacock said.<br>For example, picture air flowing around, over and past an object. &quot;Instead of flowing smoothly past the object, the air tends to dramatically part from the surface, or separate,&quot; said Peacock. Like the wake behind a boat, the water doesn't automatically reconfigure into a single stream. Rather, the region is quite turbulent. &quot;And that adversely affects the lift [or vertical forces] and drag [or horizontal forces] of the object.&quot;<br>In 1904, Ludwig Prandtl derived the exact mathematical conditions for flow separation to occur. But his work had two major restrictions: first, it applied only to steady flows, such as those around a car moving at a constant low speed. Second, it only applied to idealized two-dimensional flows.<br>&quot;Most engineering systems, however, are unsteady. Conditions are constantly changing,&quot; Haller said. &quot;For example, cars accelerate and decelerate, as do planes during maneuvers, takeoff and landing. Furthermore, fluids of technological interest really flow in our three-dimensional world,&quot; he added.<br>As a result, ever since 1904 there have been intense efforts to extend Prandtl's results to real-life problems, i.e., to unsteady three-dimensional flows.<br>A century later, Haller led a group that did just that. In 2004 Haller published his first paper in the Journal of Fluid Mechanics explaining the mathematics behind unsteady separation in two dimensions. This month, his team reports completing the theory by extending it to three dimensions. Haller's coauthors are Amit Surana, now at United Technologies; MIT student Oliver Grunberg; and Gustaaf Jacobs, now on the faculty at San Diego State University.<br>Equally important, this month Peacock and colleagues report important experimental work. Said Peacock, &quot;while we fully trust George's new mathematical results, the engineering community is usually skeptical until they also see experimental results.&quot; Haller added, &quot;while giving a beautiful validation of the 2D theory, Tom's work also gives strong experimental backing to our new 3D theory.&quot; Coauthors on the experimental work are Haller, Jacobs, Matthew Weldon, now at Penn State; and Moneer Helu, now at the University of California at Berkeley.<br>The research was initially supported by an internal source, the MIT Ferry Fund. Currently the work is supported by the Air Force Office of Scientific Research and the National Science Foundation. <br>The researchers said it's too soon to quantify the level of improvement in performance of cars and planes that might stem from the work, noting that more work must be done before it can be applied to commercial technologies. &quot;This is the tip of the iceberg, but we've shown that this theory works,&quot; Peacock said. <!--v:3.2--> ]]></description>
<category><![CDATA[工程技术]]></category>
<author><![CDATA[550115938@qq.com(wildyouth)]]></author>
<comments>http://550115938.qzone.qq.com/blog/1223141046#comment</comments>
<qz:effect>512</qz:effect>
<pubDate>Sat, 04 Oct 2008 17:24:06 GMT</pubDate>
<guid>http://550115938.qzone.qq.com/blog/1223141046</guid>
</item>

<item>
<title><![CDATA[献给所有的艺术朋友--分形与分形艺术]]></title>
<link>http://550115938.qzone.qq.com/blog/1211456733</link>
<description><![CDATA[分形与分形艺术 <br>我们人类生活的世界是一个极其复杂的世界，例如，喧闹的都市生活、变幻莫测的股市变化、复杂的生命现象、蜿蜒曲折的海岸线、坑坑洼洼的地面等等，都表现了客观世界特别丰富的现象。基于传统欧几里得几何学的各门自然科学总是把研究对象想象成一个个规则的形体，而我们生活的世界竟如此不规则和支离破碎，与欧几里得几何图形相比，拥有完全不同层次的复杂性。分形几何则提供了一种描述这种不规则复杂现象中的秩序和结构的新方法。 <br>一、分形几何与分形艺术 <br>什么是分形几何？通俗一点说就是研究无限复杂但具有一定意义下的自相似图形和结构的几何学。什么是自相似呢？例如一棵苍天大树与它自身上的树枝及树枝上的枝杈，在形状上没什么大的区别，大树与树枝这种关系在几何形状上称之为自相似关系；我们再拿来一片树叶，仔细观察一下叶脉，它们也具备这种性质；动物也不例外，一头牛身体中的一个细胞中的基因记录着这头牛的全部生长信息；还有高山的表面，您无论怎样放大其局部，它都如此粗糙不平等等。这些例子在我们的身边到处可见。分形几何揭示了世界的本质，分形几何是真正描述大自然的几何学。 <br>“分形” 一词译于英文Fractal，系分形几何的创始人曼德尔布罗特（B.B.Mandelbrot）于1975年由拉丁语Frangere一词创造而成，词本身具有“破碎”、“不规则”等含义。Mandelbrot研究中最精彩的部分是1980年他发现的并以他的名字命名的集合，他发现整个宇宙以一种出人意料的方式构成自相似的结构（见图1）。Mandelbrot 集合图形的边界处，具有无限复杂和精细的结构。如果计算机的精度是不受限制的话，您可以无限地放大她的边界。图2、图3 就是将图1中两个矩形框区域放大后的图形。当你放大某个区域，它的结构就在变化，展现出新的结构元素。这正如前面提到的“蜿蜒曲折的一段海岸线”，无论您怎样放大它的局部，它总是曲折而不光滑，即连续不可微。微积分中抽象出来的光滑曲线在我们的生活中是不存在的。所以说，Mandelbrot集合是向传统几何学的挑战。 <br><wbr /><a href="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic01.gif" target="_blank"><img style="border:0;" src="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic01.gif" /></a><wbr />图 1 Mandelbrot集合<wbr /><a href="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic02.gif" target="_blank"><img style="border:0;" src="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic02.gif" /></a><wbr />图 2 Mandelbrot集合局部放大<wbr /><a href="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic03.gif" target="_blank"><img style="border:0;" src="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic03.gif" /></a><wbr />图 3 Mandelbrot集合局部放大 <br>用数学方法对放大区域进行着色处理，这些区域就变成一幅幅精美的艺术图案，这些艺术图案人们称之为“分形艺术”。“分形艺术” 以一种全新的艺术风格展示给人们， 使人们认识到该艺术和传统艺术一样具有和谐、对称等特征的美学标准。这里值得一提的是对称特征，分形的对称性即表现了传统几何的上下、左右及中心对称。同时她的自相似性又揭示了一种新的对称性， 即画面的局部与更大范围的局部的对称，或说局部与整体的对称。 这种对称不同于欧几里德几何的对称，而是大小比例的对称，即系统中的每一元素都反映和含有整个系统的性质和信息。这一点与上面所讲的例子：“一头牛身体中的一个细胞中的基因记录着这头牛的全部生长信息”， 完全吻合。 不管你是从科学的观点看还是从美学的观点看，她都是那么富有哲理，她是科学上的美和美学上的美的有机结合。 <br>二、复平面中的神奇迭代 <br>Mandelbrot集合是Mandelbrot在复平面中对简单的式子 Z &lt;- Z^2 + C 进行迭代产生的图形。虽然式子和迭代运算都很简单，但是产生的图形出现那么丰富多样的形态及精细结构简直令人难以置信以至于不可思议。在传统几何学中难以找到如此简单的规律隐藏着如此复杂而生动的例子。Mandelbrot集合告诉我们自然界中简单的行为可以导致复杂的结果。例如，大型团体操中每个人穿的衣服只有几种颜色中的一种，每个人的动作也只是导演规定的几种之一。但是整体上可以显示出多种多样的复杂形态。 <br>Julia 集合 <br>在复平面上，水平的轴线代表实数，垂直的轴线代表虚数。每个Julia集合（有无限多个点）都决定一个常数C，它是一个复数。现在您在复平面上任意取一个点，其值是复数Z。将其代入下面方程中进行反复迭代运算： <br>　　<wbr /><a href="http://www.fractal.net.cn/aboutfractal/aboutfractal/math-1.gif" target="_blank"><img style="border:0;" src="http://www.fractal.net.cn/aboutfractal/aboutfractal/math-1.gif" /></a><wbr /> <br>就是说，用旧的Z自乘再加上C后的结果作为新的Z。再把新的Z作为旧的Z，重复运算。当你不停地做，你将最后得到的Z值有3种可能性： <br>　　1、Z值没有界限地增加（趋向无穷） <br>　　2、Z值衰减（趋向于零） <br>　　3、Z值是变化的，即非1或非2 <br>趋向无穷和趋向于零的点叫定常吸引子，很多点在定常吸引子处结束，被定常吸引子所吸引。非趋向无穷和趋向于零的点是“Julia集合”部分，也叫混沌吸引子。 <br>问题是我们怎样才能让计算机知道哪一个点是定常吸引子还是“Julia集合”。一般按下述算法近似计算： <br>　　n=0; <br>　　while ((n++ &lt; Nmax) &amp;&amp; (( Real(Z)^2 + Imag(Z)^2) &lt; Rmax)) <br>　　{ <br>　　　Z=Z*Z+C; <br>　　} <br>其中：Nmax为最大迭代次数 <br>　　　Rmax为逃离界限 <br>退出while循环有两种情况，第一种情况是： <br>　　(Real(Z)^2 + Imag(Z)^2) &gt;= Rmax <br>属于这种情况的点相当于“Z值没有界限地增加（趋向无穷）”，为定常吸引子,我们把这些区域着成白色。第二种情况是： <br>　　n &gt;= Nmax <br>属于这种情况的点相当于“Z 值衰减（趋向于零）”或“Z 值是变化的”，我们把这些区域着成黑色。黑色区域图形的边界处即为“Julia集合”。“Julia集合”有着极其复杂的形态和精细的结构。 <br>黑白两色的图形艺术感染力不强。要想得到彩色图形，最简单的方法是用迭代返回值n来着颜色。 要想获得较好的艺术效果，一般对n做如下处理： <br>　　Red = n*Ar+Br; <br>　　Grn = n*Ag+Bg; <br>　　Blu = n*Ab+Bb; <br>　　if ((Red &amp; 0x1FF) &gt; 0xFF) Red = Red ^ 0xFF; <br>　　if ((Grn &amp; 0x1FF) &gt; 0xFF) Grn = Grn ^ 0xFF; <br>　　if ((Blu &amp; 0x1FF) &gt; 0xFF) Blu = Blu ^ 0xFF; <br>其中：Ar、Ag、Ab及Br、Bg、Bb为修正量 <br>获得的Red、Grn、Blu为RGB三基色，着色效果为周期变化，具有较强的艺术感染力，而且等位线也蕴藏在周期变化的色彩之中。 <br>你可以想象得出，在屏幕上顺序的试用每个像素点来反复迭代方程要花费很长的时间。一幅 1024x768 屏幕尺寸的画面有786432个点。其中一些点在计算机上要反复迭代方程次数达1000次（取决于Nmax的取值）或更多次才放弃运算。运算产生一幅Julia集合需要花费很长的时间，有时需要产生一幅做海报用的大图像时，如 10240x7680，要花几天的时间。当然，你使用高速计算机会缩短这个时间。图 4、5、6是三幅Julia集合： <br><wbr /><a href="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic04.jpg" target="_blank"><img style="border:0;" src="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic04.jpg" /></a><wbr />图 4 象尘埃一样的结构<wbr /><a href="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic05.jpg" target="_blank"><img style="border:0;" src="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic05.jpg" /></a><wbr />图 5 稳定的固态型<wbr /><a href="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic06.jpg" target="_blank"><img style="border:0;" src="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic06.jpg" /></a><wbr />图 6 象树枝状 <br>Mandelbrot 集合 <br>将Mandelbrot集合和Julia集合联系在一起，Julia集合有若干类型，都包含在Mandelbrot集合之中。Julia集合中的C是一个常量，而Mandelbrot集合的C是由进入迭代前的Z值而定。迭代结果，Z值同样有3种可能性，即： <br>　　1、Z值没有界限地增加（趋向无穷） <br>　　2、Z值衰减（趋向于零） <br>　　3、Z值是变化的，即非1或非2 <br>Mandelbrot集合是所有的朱莉娅集合的合并，Mandelbrot集合的某个区域放大后就是这个点的Julia集合。 Mandelbrot集合有着一些很异国情调并且古怪的形状（见图1）。你能不停地永远放大Mandelbrot集合，但是受到计算机精度的限制。 <br>Newton/Nova 分形 <br>Newton奠定了经典力学、光学和微积分学的基础。但是除了创造这些自然科学的基础学科外，他还建立了一些方法，这些方法虽然比不上整个学科那么有名，但已被证明直到今天还是非常有价值的。例如，牛顿建议用一个逼近方法求解一个方程的根。你猜测一个初始点，然后使用函数的一阶导数，用切线逐渐逼近方程的根。如方程 Z^6 + 1 = 0有六个根，用牛顿的方法“猜测”复平面上各点最后趋向方程的那一个根， 你就可以得到一个怪异的分形图形。 和平常的Julia分形一样，你能永远放大下去，并有自相似性。 牛顿分形图形中的颜色显示每个答案的种类及性质， 即迭代到目的地花费的时间，如图7所示： <br><wbr /><a href="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic07.jpg" target="_blank"><img style="border:0;" src="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic07.jpg" /></a><wbr />图7 Newton分形 <br>Paul Derbyshire研究牛顿分形图形时，他把Julia集合的常值C加入进去改变了一下算法，并用同样的方法去估算Z，逼近答案，产生奇特的并称之为“Nova”的分形图形。“Nova”类型分形图形如图8所示： <br><wbr /><a href="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic08.jpg" target="_blank"><img style="border:0;" src="http://www.fractal.net.cn/aboutfractal/aboutfractal/pic08.jpg" /></a><wbr />图 8 Nova分形 <br>三、关于分形艺术的争论 <br>把计算机产生的图形看成是艺术，有人可能要提出一些疑问。这些图形可以利用高品质的打印机产生任意多幅同样质量的“原作”，从而在商业化的艺术市场上造成混乱，因此她没有收藏价值，没有收藏价值的作品还能算得上是艺术吗？ <br>这是一个十分敏感的问题。早在六十年代初有些数学家和程序设计人员就开始利用计算机及绘图设备从事这方面的工作。但他们大部分人避免将自己的工作与“艺术”一词挂起钩来，以免与艺术界的人们发生冲突。但是有一些人还是挺着腰杆去面对批评，承认计算机是视觉艺术的一种新工具，称他们自己的方法为“计算机艺术”。在批评面前，他们没有受到影响。他们不顾理论界的反对而继续自己的探索。　他们积累了大量令人难忘的成果。　正因为他们的努力才出现了今天的PhotoShop、Corel DRAW等等著名的软件，以及各种计算机艺术团体组织。PhotoShop也成了某些美术专业学生的必修课。 <br>当今时代出现的充满科技含量的“分形艺术”又不同于运用PhotoShop从事的计算机艺术创作。“分形艺术”是纯数学产物，是否能算得上艺术必然会引起新的争论。争论最活跃的问题是：分形图形是纯数学产物能算得上艺术吗？既然学习数学和程序设计就可以从事艺术创作了，学习美术专业还有什么用处呢？ <br>这个问题提的好。　从事分形艺术创作的人要研究产生这些图形的数学算法，这些算法产生的图形是无限的。他们没有结束，你永远不能看见它的全部。你不断放大她们的局部，也许你可能正在发现前人没曾见到过的图案。这些图案可能是非常精彩的。她们与现实世界相符合，从浩瀚广阔的宇宙空间到极精致的细节，是完全可以用数学结构来描述的。另一个的问题是颜色，好的颜色选择，就可以得到一幅奇妙的图形。糟糕的选择，你得到的就是垃圾。所以说，创造分形艺术，最好再学一点绘画基础、色彩学等，那将是大有益处。 <br>分形几何冲击着不同的学术领域，她在艺术领域显示出非凡的作用。创作精美的分形艺术是国内外分形艺术家们的人生追求，总有一天分形艺术会登上大雅艺术殿堂。 <br><wbr /><a href="http://www.fractal.net.cn/title/dot.gif" target="_blank"><img style="border:0;" src="http://www.fractal.net.cn/title/dot.gif" /></a><wbr /> 补充说明用上述算法，就可以创作出很多精美的图片，比如下图就是按这种算法计算出来的分行艺术作品。当然了，使你的作品更神奇、漂亮，你必需不断改进你的算法。坚持持续多年，肯定会积累很多好的算法主意。一个人能坚持数年认真做一件事情，肯定能做的很好、或很出色。 <br>多年程序设计告诉了我，一个具有实用价值的分形艺术软件，真正核心算法所占的程序篇幅并不是很大，而辅助程序所占的程序代码量非常庞大。比如：交互界面的安排、设计；表达试的编译运行；每幅分形图案参数的读取、保存、交互修改等等。这些你都要认真思考去合理设计才能使你的程序使用起来得心应手。即使你实现上述固定一个表达式的简单演示程序，你也要做大量辅助代码工作。 <br><wbr /><a href="http://www.fractal.net.cn/aboutfractal/news/n-102.jpg" target="_blank"><img style="border:0;" src="http://www.fractal.net.cn/aboutfractal/news/n-102.jpg" /></a><wbr /> <!--v:3.2--> ]]></description>
<category><![CDATA[自然哲学]]></category>
<author><![CDATA[550115938@qq.com(wildyouth)]]></author>
<comments>http://550115938.qzone.qq.com/blog/1211456733#comment</comments>
<qz:effect>513</qz:effect>
<pubDate>Thu, 22 May 2008 11:45:33 GMT</pubDate>
<guid>http://550115938.qzone.qq.com/blog/1211456733</guid>
</item>

<item>
<title><![CDATA[Applying Consumer Theory to Competitive Markets]]></title>
<link>http://550115938.qzone.qq.com/blog/1211425472</link>
<description><![CDATA[Lecture: Applying Consumer Theory to Competitive Markets<br>David Autor<br>14.03 Fall 2004<br>1 Applying Consumer Theory to Competitive Markets<br>  When exchange takes place voluntarily, economists generally assume that it makes all participants better<br>o¤. Otherwise, they would not have engaged in the exchange.<br>  It?s useful to have a dollar metric of the gains from transacting.<br>  This measure is consumer surplus.<br>  This is critical because although we can readily measure the direct costs of a given project or policy (i.e.,<br>building a bridge, imposing a tarrif), it?s less obvious how we measure the bene?ts.<br>  Demand curves allow us to measure these bene?ts.<br>  Think of market demand curve is the set of consumers arrayed in inverse ordering from the person with<br>the highest willingness to pay (WTP) for a good to the person with the lowest WTP.<br>  Similarly, think of the market supply curve as the set of producers arrayed in order from ?rm willing to<br>produce at lowest price to ?rm demanding the highest price to produce a good.<br>  What market ideally does is match consumers and producers:<br>?If a producer is willing to produce at a price less than or equal what a consumer is willing to pay,<br>that transaction will occur.<br>?Most consumers will be buying at a price below than their maximal willingness to pay.<br>?Most producters will be selling at a price above their lowest willingness to produce.<br>?Marginal producer and consumer will be indi¤erent.<br>  It is easy to see that when this mechanism works correctly, it maximizes the sum of producer and consumer<br>surplus:<br>?All gains from trade are realized.<br>?All transactions that bene?t both parties occur.<br>1<br>?No transactions occur that do not bene?t both parties.<br>  It is noteworthy that this metric does not place any greater weight on consumer or producer surplus:<br>?If supply is perfectly elastic, all of the surplus is captured by consumers.<br>?If demand is perfectly elastic, all of the surplus is captured by producers.<br>  Why do we want to maximize surplus without any regard for who are the bene?ciaries? Isn?t there a<br>tradeo¤ between equity and e￠ ciency?<br>?Answer: No, not in a competitive market. We?ll study this issue shortly in general equilibrium theory.<br>?But for the moment, take it on faith that we just want to maximize the pie, and we don?t need to<br>worry about who gets which slice.<br>  We often have pretty good data on producers?costs, which should re?ect their willingness to pay.<br>  We often have less meaningful data on consumer?s willingness to pay.<br>  But if we have an estimate of the elasticity of demand, this allows us to extrapolate a great deal of<br>information about consumer surplus. That?s because the elasticity is a measure of the slope (or curvature)<br>of the demand curve.<br>  We have previously worked with compensated demand curves. For this analysis, we?ll use the uncompen-<br>sated demand curve for sugar. Why is that reasonable? Two explanations:<br>?We don?t have the compensated demand curve (not a great reason)<br>?Recall the Slutsky equation: @dx=@px = @hx=@px ?? (@dx=@I)   x. This says that the compensated<br>and uncompensated demand curves don?t di¤er much if (@dx=@I)   x is small. This will be true if the<br>income e¤ect is small or if sugar is a small share of consumers?budgets (so, there is very little change<br>in consumers?real income when sugar prices change). The latter is quite likely to be true. Hence,<br>we don?t have to worry a great deal about the di¤erence between compensated and uncompensated<br>demand in this case.<br>  We?re going to apply this reasoning to the system of U.S. Sugar Tari¤s to perform a welfare analysis.<br>  This is a partial equilibrium analysis, using the tools of consumer theory that we have studied so far (plus<br>a little producer theory from 14.01).<br>  We will leap into general equilibrium analysis in a couple of lectures.<br>  As we proceed, three points to keep in mind:<br>2<br>1. Distinguish carefully between transfers and social gains and losses.<br>?If I pay a worker $5 to do a job, that is not a $5 social gain even though I have ?generated a job.?<br>?Why? This transaction is a $5 transfer from me to the worker.<br>?If there is a gain, it is because the value that the worker produces for $5 is worth more than $5<br>to me (i.e., consumer surplus) and/or the worker?s alternative use of time was less than $5 (i.e.,<br>producer surplus).<br>2. A cost is not a bene?t.<br>?This is a restatement of the above. If I double the generosity of welfare programs in the United<br>States, what is the social cost of this? To a ?rst approximation, zero. I am simply transfering<br>money from one group of citizens to another. The social cost, if present, would only be due<br>to the deadweight loss of taxation or the distortions that free money had on labor supply (and<br>other behavior) of welfare recipients.<br>3. All costs are opportunity costs.<br>?In economic reasoning, there is no such thing as intrinsic value. The only cost of using a given<br>resource is the value of its alternative to which it could have been put.<br>?This is it?s opportunity cost.<br>?Why is water so cheap, even though it is the source of all human life? Because there is an<br>abundance of water (in Cambridge), so the opportunity cost of the last gallon of water out of<br>your tap in the morning is quite low. But if there was a shortage, the price of water could rise<br>to extremely high levels because the value of the ?rst gallon of water is quite prescious.<br>?Similarly, diamonds are expensive because at available quantities, people appear to have very<br>high value uses for them (i.e., wedding rings). If diamonds were as abundant as water, they<br>would presumbaly be as cheap (and we would use as party favors). (Why are diamond quantities<br>so limited? Probably due to cartelization of the world diamond market by De Beers rather than<br>natural scarcity.)<br>1.1 Sugar case<br>1.1.1 Analytics<br>Using basic competitive theory, we will analyze the e￠ ciency consequences of the U.S. sugar program.<br>To do this, we need to model the consumer demand curve (to assess consumer surplus) and model the supply<br>curve to assess producer surplus.<br>With these, we can consider the consequences of the quota system relative to a counterfactual case in which<br>the market was unrestricted.<br>The deviations from the competitive baseline case can be divided into three components:<br>3<br>1. Transfers. In general, quantity or price distortions will yield some transfers from consumers to producers<br>(or vice versa). These are not e￠ ciency losses (though we may still feel negatively about them).<br>2. Deadweight losses from ine￠ cienct resource allocation. As we?ll see in the sugar case, price or quantity<br>quotas may also cause production distortions whereby low cost producers are thwarted from producing<br>and high cost producers take their place. In these cases, there is of course the usual loss in consumer<br>surplus and gain in producer surplus. But there is also a deadweight loss incurred; real resources are<br>consumed by high-cost producers to make goods that low-cost producers could have made using fewer<br>resources.<br>3. Deadweight losses from foregone consumption. A price or quantity quota will generally reduce equilbrium<br>consumption below its competitive level. This implies that there are some units of the good that consumers<br>would have been willing to pay for at a price that producers would have been willing to produce at. These<br>thwarted trades are a form of deadweight loss.<br>1.1.2 Sources of sugar supply:<br>1. World sugar supply:<br>  Price: perfectly elastic at price $0:068 per pound<br>  Quantity: capped at 3:8 billions pounds<br>2. U.S. sugar producers:<br>  upward sloping from $0:068 per pound at 1 pound to $0:22 at 14 billion pounds.<br>  So Pdomestic = 0:068 + 1:09   Q, where Q is billion pounds produced.<br>3. High fructose corn syrup (HFCS) producers:<br>  Supply: $0:15 per pound for 12 billion pounds, in?nite cost thereafter (until new plants built).<br>1.1.3 Demand:<br>We are given that: elasticity of demand is ??0:30 and that domestic demand is 29 billion pounds at $0:22 per<br>pound.<br>  Use the functional form<br>Q(P) = KP??:30<br>  Why this functional form?<br>  =<br>@Q<br>@P  <br>P<br>Q<br>= ??:30KP??1:30  <br>P<br>KP??:30 = ??:30;<br>so, this functional form has a constant elasticity.<br>4<br>  Another way to see this:<br>lnQ(P) = lnK ?? :30 ln P;<br>@ lnQ<br>@ ln P<br>= ??0:30<br>  Applying in this example:<br>29 = K(22)??0:30<br>K = 29(22)0:30 = 73:3<br>Q = 73:3(22)??0:30<br>?What would quantity demanded be at the world price:<br>Q(0:068) = 73:3  <br>??<br>6:8)??0:30 <br>= 41:2 billion pounds<br>1.1.4 Accounting<br>Gains in producer surplus?<br>  Why won?t gains to producers be identical to corresponding losses for consumers?<br>?Foreign producers (Do you want to count these gains? What are the politics that makes this large<br>transfer feasible?)<br>3:8(22 ?? 6:8)=100 = 0:58 billion<br>This amount is a pure pure transfer. Why? There is no distortion involved in taking a chunk of<br>consumer surplus and handing it over to producers. Deadweight losses accrue when consumption is<br>foregone (as it will be) or production decisions are distorted.<br>?U.S. sugar producers<br>Production Costs 13:2   (6:8 + 0:5 (22 ?? 6:8)) =100 = 1:9 billion<br>Revenue 13:2(22)=100 = 2:9 billion<br>Gain in Producer surplus = 1:0 billion<br>Loss in Consumer surplus 13:2(22 ?? 6:8)=100 = 2:0 billion<br>DWL 1:0 ?? 2:0 = ??1:0 billion<br>5<br>?HFCS producers<br>Production costs 12(15)=100 = 1:8 billion<br>Revenue 12(22)=100 = 2:64 billion<br>Gain in producer surplus = 0:84 billion<br>Loss in consumer surplus 12(22 ?? 6:8)=100 = 1:82 billion<br>DWL 0:84 ?? 1:82 = ??0:98 billion<br>  Loss of consumer surplus from foregone consumption?<br>?The DWLs calculated above are from ine￠ cient resource allocation ?expensive producers producing<br>what cheap producers could have produced were it not for the sugar program.<br>?There is also a second DWL that should look more familiar. A considerable quantity of sugar is not<br>consumed at 22 cents per pound that would have been consumed at 6:8 cents per pound.<br>?Quantify this DWL: Willingness to pay ?? production costs in the competitive case.<br>?What were consumers willing to pay for this sugar? Invert the demand curve and integrate:<br>Q(P) = 73:3P??0:3<br>P(Q) =<br> <br>Q<br>73:3<br> <br>?? 1<br>0:3<br>Z 41:2<br>29<br> <br>Q<br>73:3<br> <br>?? 1<br>0:3<br>@Q = 2:03<br>?We calculated above that at 6:8 cents, they would have consumed 41:2 billion pounds, instead of the<br>29 billion pounds consumed under the quota system. So, that?s 12:2 billion pounds of consumption<br>foregone<br>?12:2 billion pounds would have had a real cost of 0:83 billion dollars.<br>?So, the DWL of the foregone consumption if 0:83 ?? 2:03 = 1:2 billion in DWL.<br>1.2 Accounting<br>Accounting for costs and bene?ts<br>6<br>Quantity Production Cost Revenue Producer Consumer DWL<br>Produced Surplus Surplus<br>Foreign 3:8 bil 3:8(:068) =<br>0:260 bil<br>3:8(:22) =<br>0:84 bil 0:58 bil<br>3:8  (:22 ?? :068)<br>= ??0:58 bil<br>0:0 bil<br>HFCS 12 bil 12(:15) =<br>1:8 bil<br>12(:22) =<br>2:64 bil 0:84 bil<br>12  (:22 ?? :068)<br>= ??1:82 bil ??0:98 bil<br>Domestic<br>Prod?n 13:2 bil<br>13:2<br> 1<br>2 (:22 ?? :068)<br>+13:2   :068<br>= 1:90 bil<br>13:2(:22) =<br>2:9 bil 1:00 bil<br>13:2  (:22 ?? :068)<br>= ??2:00 bil ??1:00 bil<br>Reduced<br>Consump 12:2 bil 0:83 billion<br>(if produced)<br>12:2(:068) =<br>0:83 bil<br>(if prod)<br>0 (if prod)<br>:83 ?? 2:03<br>= ??1:2 bil<br>(if con-<br>sumed)<br>??1:20 bil<br>Total +2:42 bil ?? 5:60 bil ??3:18 bil<br>Notice the following identities:<br>  Change in Producer surplus = Change in revenue minus change in production costs<br>  Change in Consumer surplus = Quantity consumed times price change [This only applies to the units<br>consumed ?not the thwarted consumption]<br>  DWL for units consumed = Gains to producers - Loss to consumers<br>  DWL from thwarted consumption = Willingness to pay for foregone consumption minus cost of producing<br>those goods<br>Amazingly, the deadweight loss of the program is 2:5 times as large as the transfer to producers: 6:3 versus<br>2:4 billion. For every $1 in surplus that producers gain from the sugar quotas, consumers lose $2:5.<br>Notice that about 25 percent of the gains (0:58 of 2:42 billion) is a pure transfer to foreign producers. The<br>gain to U.S. producers is $1:84 billion.<br>Moreover, about $0:98 is a transfer not to sugar producers but to HFCS producers.<br>The transfer to domestic sugar producers is only $0:86 billion, less than 15 percent of the loss to consumers.<br>7<br>1.3 Consider<br>  At the time the case was written, there were 8; 360 sugar farms in the U.S. So this is implicitly a subsidy<br>of<br>(0:86 billion)=8; 360 = $102; 000 per farm.<br>  Also consider that 1; 400 farms account for about 50% of all production.<br>(0:86 billion x 0:5)=1400 = $307; 000 subsidy per farm for these farms.<br>  And these are the gains per farm. If we counted the loss to consumers per farm, it would be more than<br>twice as large.<br>  What would be the reaction of farmers if we proposed to open the U.S. to world sugar prices and gave a<br>cash payment of $102; 000 per (former) sugar farm per year?<br>  How would the High Fructose Corn Syrup producers (e.g., Archer Daniels Midland) feel about this?<br>1.4 What about Michael Warner of the American Sugarbeet Growers Associa-<br>tion?<br>  He points out the $300 per pound loss in domestic sugar production in North Dakota over 1<br>4 million acres?<br>Is this loss $300   250; 000?<br>  What about the 30; 000 sugar beet farming jobs in North Dakota that Warner stresses? What are the<br>economic losses if these workers are not needed for sugar production? Are they 30; 000 AnnualEarnings?<br>  Is it signi?cant that Mr. Warner counts both the loss in crop production of $300=acre and the loss of jobs<br>of 30; 000 workers as costs of eliminating the program?<br>?This is clearly double-counting.<br>?The workers should be viewed as costs to sugar producers and their wages should be subtracted from<br>the crop production per acre ?along with all other variable costs and the opportunity cost of the<br>land ?to get a measure of the potential surplus generated by using this land for sugar production.<br>?Stated di¤erently, farmers would be delighted to ?re all of their workers if they could still get $300<br>per acre in direct subsidies. So clearly you can?t count both the sugar income and the expense of<br>employing the workers as social bene?ts.<br>8 <!--v:3.2--> ]]></description>
<category><![CDATA[经济管理]]></category>
<author><![CDATA[550115938@qq.com(wildyouth)]]></author>
<comments>http://550115938.qzone.qq.com/blog/1211425472#comment</comments>
<qz:effect>512</qz:effect>
<pubDate>Thu, 22 May 2008 03:04:32 GMT</pubDate>
<guid>http://550115938.qzone.qq.com/blog/1211425472</guid>
</item>

<item>
<title><![CDATA[熵与信息]]></title>
<link>http://550115938.qzone.qq.com/blog/1211425224</link>
<description><![CDATA[位<br>就像长度以米为量度，时间以秒为量度一样，信息以位为量度。当然，知道了信息量并不等同于知道了信息本身，它是什么意思或者它意味着什么。讲义中我们不会考虑信息的内容或含义，只考虑信息量。<br>不同的环境下需要不同的尺度衡量长度。有时候我们用千米量度长度，有时侯用英尺，还有时候用埃。与之类似，除了位以外我们有时候需要其他信息的尺度；在物理系统中信息通常以焦耳每绝对温度衡量。<br>信息是怎么被量化的？考虑一种有几种可能结果的情形。举个例子，像掷一枚硬币（两种结果，正面或反面）或是从一副扑克牌中选一张拍（52种可能的结果）。一个人（按照惯例叫做Alice）怎么才能简洁地把这个结果告诉另外一个人（Bob）呢？<br>首先考虑掷一枚硬币的两种结果的情况，假设它们的可能性相等。如果Alice要告诉Bob掷硬币的结果，她有几种可能的方法，说“正面”或“反面，或者那种情况下说0或1；但根据传递的信息量，它们是一样的。我们说传递的信息是1位。<br>如果Alice掷两枚硬币，她可以说两次0或1，这样就说明了四种实际可能发生的情况。相似地，一个8种可能结果的试验的结果可以用3位来表达；更一般地，种结果的试验用n位表达。这样，信息量是可能结果的对数（以2为底）。 n2<br>注意传递信息需要两个阶段。第一个为“起始”阶段，在这个阶段Alice和Bob对他们将要传递什么信息，每个位究竟代表什么达成一致的意见。这样共有的认识称为编码。这个过程在结果产生以前就完成了。然后，就是“通讯”阶段，在这个阶段发送0和1的序列。这些序列叫做数据。所以要表达一副拍的花色，编码可以是00代表梅花，01代表方块，10代表红桃，11代表黑桃。用这个编码Alice就可以抽一张拍，然后发送两位数据来告诉Bob它的花色。用同样的编码，她可以重复地做多个实验。<br>在Bob知道有一张拍被抽之后，但收到Alice的信息之前，他并不确定花色。这个不确定性，或信息缺失也可以用位来表达。听到结果，他的不确定性减少了，因为他收到了信息。Bob的不确定性在起始阶段增加，在通讯阶段减少。<br>注意一些关于信息的重要结论，其中一些在本例中已经说明。<br>?? 通过观察，实验和测量可以得到信息。<br>?? 信息是主观的，或者说“依赖于观察者的”。Alice知道的东西和Bob知道的东西不同。（如果信息不是主观的，也就不需要通讯来传递信息了。）<br>?? 当一个人知道了在一次观察中信息是有效的，他对事情的不确定性就增加了，当收到了那条信息，不确定性就减少了。<br>?? 可以丢失信息，或者通过丢失数据本身，或者丢失编码。<br>?? 用空间和时间可以定位信息的物理形式。如一个序列……<br>－信息可以从一个地方被发送到另一个地方。<br>－信息可以被储存，并在之后重新获得。<br>1.1 数学位<br>我们看到，信息可以通过01序列来传递。这一非常有力的抽象使得我们可以忽略特定信息处理和传输系统中伴随的许多细节。<br>位很简单，只有两个可能的值，而且用于处理单个位的数学并不难。以数学家George Boole（1815-1864）的名字命名，这就是布尔代数。在某些方面，布尔代数与高中讲授的整数或实数的代数相似，但在某些方面又不同。<br>代数涉及的是有某些可能值的变量，还涉及当给出一个或更多的变量时会返回某些可能值的函数。在布尔代数中可能值是0和1。<br>准确的说，有4种单变量的布尔函数。其中一个称为单位函数，它只是简单地返回参数。另一个称为否定函数，或负函数、反函数、互补函数，它把0变为1，反之亦然。另外2个函数不管参数是什么都返回0或1，。描述这些函数最简单的方法就是给出一个带有运算结果的表格：<br>表1.1：单变量布尔函数<br>我们注意到布尔函数比涉及整数和实数的代数简单得多，无疑这些代数都有很多单变量函数。<br>对两个变量A和B有多少个布尔函数呢？这2个变量中，每一个都可能有2个值，所以有4种可能的输入组合。将2个布尔值赋给4个输入将会有16种不同的方法。这16种方法中，2个与输入无关，4个将输出赋予A或B，或是它们的相反的值，其他10个取决于2个输入参数。最常用的是与，或，异或，与非和或非，如表1.2所示。<br>我们往往会将布尔值0和1看作和整数0和1是一样的。那么从某种程度上说，与就是相乘，或就是相加。然而，简单地从一般代数得到的相似的结果对布尔代数并不成立，所以这样的类比是危险的。尽管有时候不是很容易，但是完全有必要区分整数和布尔值。<br>表1.2：双变量布尔函数<br>布尔代数所用的标准符号使得这一任务变得更难了。（尽管它可能引起混淆，但在这儿我们仍使用这个符号，因为实际当中难以使用不至引起混淆的符号。）与函数表示为乘法的方式，将两个布尔值相邻排列，中间加一点：A与B写成A·B。或函数用加号表示：A+B就是A或B。取反，或称非函数，表示为在符号或表达式上边加一杠，这样取反A就为A。最后，异或函数XOR表示为内部含有加号的一个圆圈，BA⊕。<br>布尔函数有些一般的性质非常有用。通常可以简单地通过对所有有限的可能组合的数进行示范运算进而证明这些性质。<br>取反 A<br>与 BA?<br>与非 BA?<br>或 BA+<br>异或 BA⊕<br>表1.3：布尔逻辑符号<br>一个函数如果输出已知可以求出输入，则称该函数是可逆的。由此4个单变量函数中2个是可逆的（事实上也是自逆的（self-inverse））。两个（或更多）输入的函数中显然都是不可逆的，因为输入变量多于输出变量，但是两个或两个以上这样的函数的组合可能是可逆的；比如可以很容易证明当异或函数BA⊕乘以返回第一个参数的函数时是可逆的。<br>对于两个变量的函数来说，有很多需要考虑的性质。比如设A，B和C是布尔变量，或为0或为1，那么函数与是可交换的，因为ABBA?=?。16个函数中很多也是可交换的。表4.4概括了布尔代数的一些性质。<br>表1.4：布尔代数的性质<br>布尔代数还用到一些符号。这里使用的是最常见的。有时候与，或和非表示为，和。有时候使用中缀符号，，，),(BAAND),(BAOR)(ANOTBABA?∧表示BABA+∨表示AA表示~。布尔代数在数学逻辑中也非常有用，其中通常地符号，，BABA?∧表示BABA+∨表示AA表示?。<br>1.2 物理位<br>如果信息可以被存储或被传输，它一定有一种物理形式。存储位的装置一定有两个状态，<br>一个表示0，另一个表示1。将装置设置为这些状态中的一个或另外一个，这样就存储了一位。当需要该位时就测量该装置的状态。如果这个装置从一个地方被移动到另外一个地方，那么就发生了通讯。如果该装置持续存在一段时间，那么就可以把它当作存储器。如果该装置随机地改变其值，那么它便丢失了初始值。<br>我们通常对小的物理设备感兴趣。一个存储一位信息的物体有多小的极限来自于量子力学。一个量子位是一个可以存储一位信息的物体，但它非常小以至从属于量子力学对测量定义的极限。特别地，如果不能改变一个被测量的物体，那就不能进行测量。另外一方面，如果很多个相互作用的物体代表一位，那就可以进行测量，而且有足够多的物体可以保持不变，这样就可以再次测量相同的位。如果不是所有的物体都用相同的方式测量，那么结果可能处于两种可能布尔值的中间状态。<br>1.3 标准位<br>在当今的电子系统中许多设备可以携带信息，所有设备都采用同样的方式（或至少是一种方便查看的方式）。这样半导体存储器用可能一千个电子的聚集或缺失来存储一位。相似地无线电通信中使用了大量的光子。<br>因为涉及大量的设备，对它们的测量不能仅局限于简单的是或不是，而应该是在一段连续值之间变化。所以一个半导体逻辑元件的电压可能就是在一个范围，比如0V到5V，之间的任何值。电压值也允许存在小范围的误差，所以0V到1V之间的电压都表示逻辑0，4V到5V之间的电压都表示逻辑1。电路也许不会保证对1V到4V之间的电压做出合适的说明。<br>如果电路中的噪声总小于1V，那么每个门电路的输出或为0V或为5V，这样电压总是表示为没有误差的位。因为信息处理时消除了小范围误差，这种电路显示的值被称为“恢复逻辑”。现代计算机的鲁棒性依靠恢复逻辑的应用。<br>标准位是一个可以无干扰地测量一个位的抽象概念。所以可以得到标准位的副本（量子位不可以被复制）。这个抽象概念对应用恢复逻辑的电路非常有用。<br>因为所有的物理系统最终将遵循量子力学理论，标准位也只是对实际情况的一个理想化近似。然而，即使是对于最先进最小的可用设备，这也是一个应用地极好的概念。<br>有趣的问题是随着半导体技术减小组件的尺寸，标准位的概念是否将继续有用。最终，当我们试图控制涉及少量原子或光子的位时，量子力学的限制作用将变得非常重要。很难准确地确定何时会发生这些情况，但某些人相信在2015年以前它就会发生。<br>1.4 量子位<br>根据量子力学，一个小的物体可能有两种可测量的两种状态。这听上去非常适合存储位。但是如果这两种状态有相同的能量，那么有可能预备一个物体，它组合了这两种状态。那还能测量它么？<br>在标准的环境中，测量可以精确地确定这样的组合是多少。此外，为了更准确的测量需要重复测量，然后取各个结果的平均值。但是，量子环境不同。量子测量中，有一个问题就是一个物体是否会保持某个特定的状态，回答总是“是”或“不是”，从来不是“可能”，比如，从来不会是“27％是，73不是。”此外，根据回答，在测量后系统会结束一个状态，这样以后的测量不会产生附加的信息。不能预测所有特定测量的结果，但是根据概率有两种可能的答案。量子力学这个奇特的性质带来了单个量子位能携带的信息的极限，也带来了一个<br>设计出可以利用这些特性的系统的机会。<br>我们将通过一个例子来说明量子位。假设量子为一个光子，它是电磁辐射（包括无线电，电视，和光）的初等粒子。光子是一个很好的将信息从一个地方带向另外一个地方的备选对象。它很小，运行很快。<br>光子有同时震荡着的电场和磁场。电场的方向称为极化（在此我们不考虑圆形极化光子）。如果一个光子朝向z轴运动，它的电场则可能在x轴运动，可能在y轴运动，或者事实上在x-y平面（水平－垂直平面）的任何一个方向上。<br>极化作用可以用来存储一位信息。如果该位是0，Alice就可以准备一个水平极化的光子，如果位是1，可以准备一个垂直极化光子。然后当Bob得到光子时，他可以测量它的垂直极化性（比如，问问极化作用是否是垂直方向的）。如果答案是“是”，那他就会推断出该位是1。<br>可能会想到用单个光子的极化作用传递多于一位的信息。为什么不让Alice用不同于水平和垂直的角度发送两位信息？为什么不用水平的，垂直的，向右侧倾斜的中间角度和向左侧倾斜的中间角度。问题在于Bob不得不决定测量什么角度。由于量子力学的限制，他不能问“极化的角度是多少”，但只能问“极化作用是不是在我选择测量的那个方向。”他测量的结果也只能是“是”或“不是”，换句话说，也只能是一个单个位。测量后，光子或者竖在他测量的那个平面内（如果结果是“是”），或者垂直于它（如果结果是“不是”）。<br>如果Bob想更精确地测量极化角度，为什么不让他多次重复测量，然后取平均值呢？这不管用，因为他的第一次测量复原了他要测量的极化角度，或者使其垂直于这个角度。这样随后的测量也都完全相同。<br>或者Bob想得到光子的多个副本，然后一一测量。这个方法同样也不凑效。他只能通过测量其性质而复制一个光子，然后用完全相同的性质创造一个新光子。他创造的所有光子都是一样的。<br>如果Alice准备了一个任意角度的光子，那Bob会测量出什么呢？或者更进一步说，如果光子在路途过程中受到随即影响改变了其极化角度呢？或者如果光子被一个邪恶偷听者在某个角度测量，然后因此而恢复到那个角度呢？在这些情况下，不管他选择什么样的极化角度，Bob总会得到“是”或“不是”的答案，更接近实际极化的角度更可能朝向答案是“是”的那个方向。更明确的说，肯定答案的概率是Bob测量角度和Alice准备发送的角度之间的那个角度的余弦的平方。不可能预测Bob的任何一个测量结果。这种本质上的随机性就是量子力学的一个概貌。<br>到现在为止还未提到：在某种方式下当准备发送两个或两个以上的量子时，量子位还有其他有趣的性质。其中一个性质（我们现在不准备讨论）被称为牵连性，它使得去向不同地方的两个光子有彼此相关的单一状态，这样对其中一个的测量会影响到随后的对另外一个光子的测量。<br>需要注意并不是所有的量子系统都有这里讨论的特性。标准位模型已经足够用到某些量子系统中。比如，如果限制极化角度为水平或垂直，就不会有噪声干扰，测量的适当的角度也已知，那么总可以在没有干扰该量子的情况下进行测量。在这种特殊的情况下，复制光子就是可能的。如果有个非常小的噪声，小到极化角度干扰一点也不会影响到测量的概率，那么在这个意义下就存在某个恢复逻辑，以至于在测量后极化作用总会恢复到水平的或垂直的。<br>1.5 量子位的优点<br>量子环境下可以做些并不怎么高尚的事情。其中一些是有益的。再次考虑Alice想用极<br>化的光子给Bob传递信息。她可以以任何角度准备这个光子，然后在初始阶段告诉Bob是什么角度。现在假设破坏者Sam想在Alice和Bob之间的路程中的某个位置处理该光子，以破坏此次通信。他制造了一个机器，它可以以他选择的角度反射该极化作用。这样，如果他选择45 ，每个水平的光子就变成了垂直的光子，反之亦然。知道了Alice将为编码为水平的或垂直的光子，Sam将角度设置为45 ，随即地选择该条信息中一半的光子，对其使用这个机器。<br>因为Alice发送给Bob的信息是水平和垂直编码，一半的位会被Sam改变，因为Bob测量的每个光子或者与Alice发送的相同，或者是另外一个布尔值，通信也就不可能有用了。<br>Alice知道了Sam的诡计，想要和Bob重新建立可靠的通信。她该怎么办？<br>她告诉Bob（用一条Sam无法偷听到的路径）用45 和135 测量光子。Sam的机器将极化角度反射了45 ，所以用Alice选择的两个状态的任意一个，这一反射将不会影响结果。当然，如果Sam发现了Alice这么做，他可以把机器重新旋转到垂直方向。或者还有其他可行的测量与反测量的方法。<br>这一幕反应了光子的量子性质，也反应了除了在某种特定极化角度下，单个光子不可能被Sam测量到。因此Alice阻止Sam的技术也就不可能用于标准位了。 <!--v:3.2--> ]]></description>
<category><![CDATA[工程技术]]></category>
<author><![CDATA[550115938@qq.com(wildyouth)]]></author>
<comments>http://550115938.qzone.qq.com/blog/1211425224#comment</comments>
<qz:effect>512</qz:effect>
<pubDate>Thu, 22 May 2008 03:00:24 GMT</pubDate>
<guid>http://550115938.qzone.qq.com/blog/1211425224</guid>
</item>

<item>
<title><![CDATA[中华人民共和国反洗钱法]]></title>
<link>http://550115938.qzone.qq.com/blog/1211424544</link>
<description><![CDATA[<div style="text-align:center;"><span style="font-weight:bold"><wbr />中华人民共和国反洗钱法</span><wbr /></div><div style="text-align:center;">(2006年10月31日第十届全国人民代表大会常务委员会第二十四次会议通过)</div> <br><div style="text-align:center;">中华人民共和国主席令第五十六号</div>　　《中华人民共和国反洗钱法》已由中华人民共和国第十届全国人民代表大会常务委员会第二十四次会议于2006年10月31日通过，现予公布，自2007年1月1日起施行。<br> <br><div style="text-align:right">　　                     中华人民共和国主席 胡锦涛           </div><div style="text-align:right">2006年10月31日           </div> <br>　　目录<br>　　第一章总则<br>　　第二章反洗钱监督管理<br>　　第三章金融机构反洗钱义务<br>　　第四章反洗钱调查<br>　　第五章反洗钱国际合作<br>　　第六章法律责任<br>　　第七章附则<br><br>　　第一章 总则<br>　　第一条 为了预防洗钱活动，维护金融秩序，遏制洗钱犯罪及相关犯罪，制定本法。<br>　　第二条 本法所称反洗钱，是指为了预防通过各种方式掩饰、隐瞒毒品犯罪、黑社会性质的组织犯罪、恐怖活动犯罪、走私犯罪、贪污贿赂犯罪、破坏金融管理秩序犯罪、金融诈骗犯罪等犯罪所得及其收益的来源和性质的洗钱活动，依照本法规定采取相关措施的行为。<br>　　第三条 在中华人民共和国境内设立的金融机构和按照规定应当履行反洗钱义务的特定非金融机构，应当依法采取预防、监控措施，建立健全客户身份识别制度、客户身份资料和交易记录保存制度、大额交易和可疑交易报告制度，履行反洗钱义务。<br>　　第四条 国务院反洗钱行政主管部门负责全国的反洗钱监督管理工作。国务院有关部门、机构在各自的职责范围内履行反洗钱监督管理职责。<br>　　国务院反洗钱行政主管部门、国务院有关部门、机构和司法机关在反洗钱工作中应当相互配合。<br>　　第五条 对依法履行反洗钱职责或者义务获得的客户身份资料和交易信息，应当予以保密；非依法律规定，不得向任何单位和个人提供。<br>　　反洗钱行政主管部门和其他依法负有反洗钱监督管理职责的部门、机构履行反洗钱职责获得的客户身份资料和交易信息，只能用于反洗钱行政调查。<br>　　司法机关依照本法获得的客户身份资料和交易信息，只能用于反洗钱刑事诉讼。<br>　　第六条 履行反洗钱义务的机构及其工作人员依法提交大额交易和可疑交易报告，受法律保护。<br>　　第七条 任何单位和个人发现洗钱活动，有权向反洗钱行政主管部门或者公安机关举报。接受举报的机关应当对举报人和举报内容保密。<br><br>　　第二章 反洗钱监督管理<br>　　第八条 国务院反洗钱行政主管部门组织、协调全国的反洗钱工作，负责反洗钱的资金监测，制定或者会同国务院有关金融监督管理机构制定金融机构反洗钱规章，监督、检查金融机构履行反洗钱义务的情况，在职责范围内调查可疑交易活动，履行法律和国务院规定的有关反洗钱的其他职责。<br>　　国务院反洗钱行政主管部门的派出机构在国务院反洗钱行政主管部门的授权范围内，对金融机构履行反洗钱义务的情况进行监督、检查。<br>　　第九条 国务院有关金融监督管理机构参与制定所监督管理的金融机构反洗钱规章，对所监督管理的金融机构提出按照规定建立健全反洗钱内部控制制度的要求，履行法律和国务院规定的有关反洗钱的其他职责。<br>　　第十条 国务院反洗钱行政主管部门设立反洗钱信息中心，负责大额交易和可疑交易报告的接收、分析，并按照规定向国务院反洗钱行政主管部门报告分析结果，履行国务院反洗钱行政主管部门规定的其他职责。<br>　　第十一条 国务院反洗钱行政主管部门为履行反洗钱资金监测职责，可以从国务院有关部门、机构获取所必需的信息，国务院有关部门、机构应当提供。<br>　　国务院反洗钱行政主管部门应当向国务院有关部门、机构定期通报反洗钱工作情况。<br>　　第十二条 海关发现个人出入境携带的现金、无记名有价证券超过规定金额的，应当及时向反洗钱行政主管部门通报。<br>　　前款应当通报的金额标准由国务院反洗钱行政主管部门会同海关总署规定。<br>　　第十三条 反洗钱行政主管部门和其他依法负有反洗钱监督管理职责的部门、机构发现涉嫌洗钱犯罪的交易活动，应当及时向侦查机关报告。<br>　　第十四条 国务院有关金融监督管理机构审批新设金融机构或者金融机构增设分支机构时，应当审查新机构反洗钱内部控制制度的方案；对于不符合本法规定的设立申请，不予批准。<br><br>　　第三章 金融机构反洗钱义务<br>　　第十五条 金融机构应当依照本法规定建立健全反洗钱内部控制制度，金融机构的负责人应当对反洗钱内部控制制度的有效实施负责。<br>　　金融机构应当设立反洗钱专门机构或者指定内设机构负责反洗钱工作。<br>　　第十六条 金融机构应当按照规定建立客户身份识别制度。<br>　　金融机构在与客户建立业务关系或者为客户提供规定金额以上的现金汇款、现钞兑换、票据兑付等一次性金融服务时，应当要求客户出示真实有效的身份证件或者其他身份证明文件，进行核对并登记。<br>　　客户由他人代理办理业务的，金融机构应当同时对代理人和被代理人的身份证件或者其他身份证明文件进行核对并登记。<br>　　与客户建立人身保险、信托等业务关系，合同的受益人不是客户本人的，金融机构还应当对受益人的身份证件或者其他身份证明文件进行核对并登记。<br>　　金融机构不得为身份不明的客户提供服务或者与其进行交易，不得为客户开立匿名账户或者假名账户。<br>　　金融机构对先前获得的客户身份资料的真实性、有效性或者完整性有疑问的，应当重新识别客户身份。<br>　　任何单位和个人在与金融机构建立业务关系或者要求金融机构为其提供一次性金融服务时，都应当提供真实有效的身份证件或者其他身份证明文件。<br>　　第十七条 金融机构通过第三方识别客户身份的，应当确保第三方已经采取符合本法要求的客户身份识别措施；第三方未采取符合本法要求的客户身份识别措施的，由该金融机构承担未履行客户身份识别义务的责任。<br>　　第十八条 金融机构进行客户身份识别，认为必要时，可以向公安、工商行政管理等部门核实客户的有关身份信息。<br>　　第十九条 金融机构应当按照规定建立客户身份资料和交易记录保存制度。<br>　　在业务关系存轩期间，客户身份资料发生变更的，应当及时更新客户身份资料。<br>　　客户身份资料在业务关系结束后、客户交易信息在交易结束后，应当至少保存五年。<br>　　金融机构破产和解散时，应当将客户身份资料和客户交易信息移交国务院有关部门指定的机构。<br>　　第二十条 金融机构应当按照规定执行大额交易和可疑交易报告制度。<br>　　金融机构办理的单笔交易或者在规定期限内的累计交易超过规定金额或者发现可疑交易的，应当及时向反洗钱信息中心报告。<br>　　第二十一条 金融机构建立客户身份识别制度、客户身份资料和交易记录保存制度的具体办法，由国务院反洗钱行政主管部门会同国务院有关金融监督管理机构制定。金融机构大额交易和可疑交易报告的具体办法，由国务院反洗钱行政主管部门制定。<br>　　第二十二条 金融机构应当按照反洗钱预防、监控制度的要求，开展反洗钱培训和宣传工作。<br><br>　　第四章 反洗钱调查<br>　　第二十三条 国务院反洗钱行政主管部门或者其省一级派出机构发现可疑交易活动，需要调查核实的，可以向金融机构进行调查，金融机构应当予以配合，如实提供有关文件和资料。<br>　　调查可疑交易活动时，调查人员不得少于二人，并出示合法证件和国务院反洗钱行政主管部门或者其省一级派出机构出具的调查通知书。调查人员少于二人或者未出示合法证件和调查通知书的，金融机构有权拒绝调查。<br>　　第二十四条 调查可疑交易活动，可以询问金融机构有关人员，要求其说明情况。<br>　　询问应当制作询问笔录。询问笔录应当交被询问人核对。记载有遗漏或者差错的，被询问人可以要求补充或者更正。被询问人确认笔录无误后，应当签名或者盖章；调查人员也应当在笔录上签名。<br>　　第二十五条 调查中需要进一步核查的，经国务院反洗钱行政主管部门或者其省一级派出机构的负责人批准，可以查阅、复制被调查对象的账户信息、交易记录和其他有关资料；对可能被转移、隐藏、篡改或者毁损的文件、资料，可以予以封存。<br>　　调查人员封存文件、资料，应当会同在场的金融机构工作人员查点清楚，当场开列清单一式二份，由调查人员和在场的金融机构工作人员签名或者盖章，一份交金融机构，一份附卷备查。<br>　　第二十六条 经调查仍不能排除洗钱嫌疑的，应当立即向有管辖权的侦查机关报案。客户要求将调查所涉及的账户资金转往境外的，经国务院反洗钱行政主管部门负责人批准，可以采取临时冻结措施。<br>　　侦查机关接到报案后，对已依照前款规定临时冻结的资金，应当及时决定是否继续冻结。侦查机关认为需要继续冻结的，依照刑事诉讼法的规定采取冻结措施；认为不需要继续冻结的，应当立即通知国务院反洗钱行政主管部门，国务院反洗钱行政主管部门应当立即通知金融机构解除冻结。<br>　　临时冻结不得超过四十八小时。金融机构在按照国务院反洗钱行政主管部门的要求采取临时冻结措施后四十八小时内，未接到侦查机关继续冻结通知的，应当立即解除冻结。<br><br>　　第五章 反洗钱国际合作<br>　　第二十七条 中华人民共和国根据缔结或者参加的国际条约，或者按照平等互惠原则，开展反洗钱国际合作。<br>　　第二十八条 国务院反洗钱行政主管部门根据国务院授权，代表中国政府与外国政府和有关国际组织开展反洗钱合作，依法与境外反洗钱机构交换与反洗钱有关的信息和资料。<br>　　第二十九条 涉及追究洗钱犯罪的司法协助，由司法机关依照有关法律的规定办理。<br><br>　　第六章 法律责任<br>　　第三十条 反洗钱行政主管部门和其他依法负有反洗钱监督管理职责的部门、机构从事反洗钱工作的人员有下列行为之一的，依法给予行政处分：<br>　　(一)违反规定进行检查、调查或者采取临时冻结措施的；<br>　　(二)泄露因反洗钱知悉的国家秘密、商业秘密或者个人隐私的；<br>　　(三)违反规定对有关机构和人员实施行政处罚的；<br>　　(四)其他不依法履行职责的行为。<br>　　第三十一条 金融机构有下列行为之一的，由国务院反洗钱行政主管部门或者其授权的设区的市一级以上派出机构责令限期改正；情节严重的，建议有关金融监督管理机构依法责令金融机构对直接负责的董事、高级管理人员和其他直接责任人员给予纪律处分：<br>　　(一)未按照规定建立反洗钱内部控制制度的；<br>　　(二)未按照规定设立反洗钱专门机构或者指定内设机构负责反洗钱工作的；<br>　　(三)未按照规定对职工进行反洗钱培训的。<br>　　第三十二条 金融机构有下列行为之一的，由国务院反洗钱行政主管部门或者其授权的设区的市一级以上派出机构责令限期改正；情节严重的，处二十万元以上五十万元以下罚款，并对直接负责的董事、高级管理人员和其他直接责任人员，处一万元以上五万元以下罚款：<br>　　(一)未按照规定履行客户身份识别义务的；<br>　　(二)未按照规定保存客户身份资料和交易记录的；<br>　　(三)未按照规定报送大额交易报告或者可疑交易报告的；<br>　　(四)与身份不明的客户进行交易或者为客户开立匿名账户、假名账户的；<br>　　(五)违反保密规定，泄露有关信息的；<br>　　(六)拒绝、阻碍反洗钱检查、调查的；<br>　　(七)拒绝提供调查材料或者故意提供虚假材料的。<br>　　金融机构有前款行为，致使洗钱后果发生的，处五十万元以上五百万元以下罚款，并对直接负责的董事、高级管理人员和其他直接责任人员处五万元以上五十万元以下罚款；情节特别严重的，反洗钱行政主管部门可以建议有关金融监督管理机构责令停业整顿或者吊销其经营许可证。<br>　　对有前两款规定情形的金融机构直接负责的董事、高级管理人员和其他直接责任人员，反洗钱行政主管部门可以建议有关金融监督管理机构依法责令金融机构给予纪律处分，或者建议依法取消其任职资格、禁止其从事有关金融行业工作。<br>　　第三十三条 违反本法规定，构成犯罪的，依法追究刑事责任。<br><br>　　第七章 附则<br>　　第三十四条 本法所称金融机构，是指依法设立的从事金融业务的政策性银行、商业银行、信用合作社、邮政储汇机构、信托投资公司、证券公司、期货经纪公司、保险公司以及国务院反洗钱行政主管部门确定并公布的从事金融业务的其他机构。<br>　　第三十五条 应当履行反洗钱义务的特定非金融机构的范围、其履行反洗钱义务和对其监督管理的具体办法，由国务院反洗钱行政主管部门会同国务院有关部门制定。<br>　　第三十六条 对涉嫌恐怖活动资金的监控适用本法；其他法律另有规定的，适用其规定。<br>　　第三十七条 本法自2007年1月1日起施行。<br> <!--v:3.2--> ]]></description>
<category><![CDATA[法律]]></category>
<author><![CDATA[550115938@qq.com(wildyouth)]]></author>
<comments>http://550115938.qzone.qq.com/blog/1211424544#comment</comments>
<qz:effect>512</qz:effect>
<pubDate>Thu, 22 May 2008 02:49:04 GMT</pubDate>
<guid>http://550115938.qzone.qq.com/blog/1211424544</guid>
</item>

<item>
<title><![CDATA[教育和家庭]]></title>
<link>http://550115938.qzone.qq.com/blog/1211424431</link>
<description><![CDATA[教育和家庭: 一些说明<br>本文的目标是理解教育投资，家庭决策制定与信贷约束之间的关系。<br>1 基本模型<br>生产技术: 假设社会中只有一种最终产品，但有两种人力投入——有技能的和无技能<br>的劳动力。最终产品是使用这两种技术结合投入生产出来的：<br>y = f (H, L, a)<br>其中H和d L分别是在生产中使用到的有生产技能和无生产技能的劳动力总和， 而?是描<br>述技术本质的参数。我们假设如果两种投入一起考虑的时候生产函数维持不变收益，而考<br>虑单个投入时收益递减。这个公式也提出了所有水平的人力资本都是完全替代的限制——<br>这也是上文曾经引用的Mookherjee和Ray的论文的一个观点, 与这个假设有关也引出了持续<br>不平等的一个额外来源。<br>劳动力供给: 在经济环境中，假设每个人都是一个无技能的劳动力单位。另外，他们还<br>拥有一定数量的技能单位，这与他们在人力资本中的投资一致。<br>生命周期: 经济主体的生活分为三个阶段。.在第一个阶段他们没有消费，也不会工作：<br>他们所做的一切就是获取技能。第二阶段是人们工作和挣取收入的时候。我们假设劳动没有<br>相关的不利因素。收入可以用于当期消费，或者存储或者投资于下一期他们不再有工资收<br>入时的消费。在第二阶段.，他们也会刚好生一个孩子。所以人口保持不变，假设每一个群<br>体都为一个规模。<br>人力资本：人力资本由人力资本和无技能的劳动力的组合产生。获得h单位人力资本的成<br>本是?s (h, h-)单位的人力资本和(1 - g)s(h, h-)单位的无技能劳动力, 其中0≤g≤ 1 ,and h- 是需<br>要人力资本的人的父母的人力资本。我们假设<br>其中S(h) = s(h, h).第二和第三个假设告诉我们成本函数是递增和凸的，所以排<br>除非凸状。接下来的两个假设持有的观念是家庭环境是有影响的——有技能的家长的孩子<br>发现获得技能要容易一些——通过一个递减率。最后一个假设的情况是h的增长导致的成本<br>的增长没有被下一代得到更多的教育导致教育他们的子女更容易的事实所带来的成本减少<br>所超过。<br>市场：在这篇文章中我们假设劳动和技能是在完全竞争市场中, 在时点t上劳动力的价<br>格是 ,技能的价格是.不过在下一节是个例外,我们将作一个极端的假设,那就是不存<br>在资本市场,以及除了人力资本以外没有其他资产.<br>政策: 政策工具, 我们假设存在e0 + e1E 数量的教育补贴, 其中<br>是家庭在教育上的花费.这笔资金是由对部分或全<br>部社会上挣钱的成员征税筹集到的,税收中一部分是人头税,一部分是纳税人的人力资本函数:<br>: 人头税当然是标准的,而对人力资本所得的征税是一个引进<br>再分配税收的简单方式.为了使结果更明显以及减少案例的个数,我们将主要将视线集中<br>在家庭已经在教育上花费了一部分钱,然后关注税收和补贴的小变化所带来的影响这样的案<br>例上.这使我们避免遇到角点解1的问题.<br>偏好：我们允许人们既从私人也从集体（家庭层面）成果中得到效用。私人成果包括物质<br>消费和象征性消费。物质消费是当一个人中年或者老年时对最终物品的消费: Ut M ,一个人在<br>t阶段的物质消费为Ut M (ct , pt+1) ,其中ct 是他中年阶段的消费, pt+1是他在生命最<br>后阶段的消费.象征性消费包括对孩子的一些光环性质的给予 (Andreoni, 1989), 为有一<br>个受到良好教育或有钱的孩子而感到骄傲,或者很愉悦的说自己的孩子在贵族学校读书.换<br>句话说,私人产出所带来的效用可以采取以下形式:<br>其中ht 是当代人的人力资本水平, ht+1是下一代的人力资本水平 (“我儿子有博<br>士学位” ), ct+1 是下一代的消费水平 (” 我的女儿开一辆保时捷” ,<br>时在教育上的花费 (“我儿子去了埃克塞特大学”). 因此私<br>人总效用为<br>集体所得的效用, 用来表示.<br>最后让U代表个人的总效用, 即<br>-------------------------------------------------------------------------------------------<br>1不过,这个公式包括没有办法将教育补贴和总收入补贴区分开来所造成的成本,因为即使将补贴中用于教育的<br>部分标记出来,家庭成员也总能将已经用在教育上的开支缩减.<br>这个公式相对来说是一般的,但仍有一些严格的限制.我们特别假设在不同水平下的<br>可分割性,以及排除孩子对接受福利救济的父母的关照考虑.<br>一个特例是标准的Barro-Becker的排他性公式: 这是当 c = 1, ,以及 s &gt;<br>1时的事例,我们把这个事例叫做完全利他主义.以区别于,或者s &gt; 1<br>时至少, 之一≠0的不完全利他主义的事例。一个更为特别的事例是.s = 0,m = 1时的象<br>征性消费.可以很轻易的证明这得出标准的Barro-Becker偏好, 是如下形式的效用函数:<br>两代人之间的契约: 考虑到每一代都必须依赖上一代的教育投资决策, 在这个分析中两<br>代人之间的契约可能会在潜在上起到很重要的作用. 例如,假设t 代人需要依赖 t – 1代的人力<br>资本投资.但如果t – 1代人不关心下一代人的福利,他们为什么要投资呢? 一个可能的解决方<br>法是t代人与上一代人订立契约,他们为上一代人提供老年保障(pt ),而上一代人以人力资本投<br>资(ht )作为回报.这明显将提供一些投资动机.但是,t 代人乐意提供的老年保障将取决于他们<br>得到了多少可供消费的钱ct , 而考虑到预算约束,这明显取决于他必须为下一代投资多少人力<br>资本(ht+1). 换句话说,这将取决于<br>t代人和t + 1代人的契约, 而这反过来又取决于t + 1代<br>人与t + 2代人之间的契约, 等等.<br>即使存在利他主义时，同样的问题也会产生.根本的问题是当代人对将来产出的偏好<br>可能会跟起决定权的一代人很不一样.当然,可能会在几代人之间存在是使这些问题减到最<br>少的契约(不言明的) , 但跨越很多代人时契约中明显存在着困难.<br>契约之间存在着外部性的事实表明理想的契约可能是涵盖每代人的.在这样的契约中,<br>{ } 1 0 , , t<br>t t t t p c h =<br>+ = 可能同时被选择以使0<br>t<br>t t l U = 最大化. 显然,我们不应该把它当成一个真正<br>的契约:可能它最好被认为是结合特定时代的所有代人的规范数.我们称之为完全契约.<br>这很显然是一个极端的假设.另一个可供选择的方法是走到另一个极端-就仅仅假设几<br>代人间的契约是不可能的,即没有契约. 则每代人都会在面临一系列的决定时被委派控制权.<br>我们添加一个合理的限制,即这一系列决定中不包括他们过世后采取的行动,我们将t代人<br>的潜在控制集定义为{ } 1 1 1 1 1 2 , , , , , , , , t t t t t t t t t t D p p p c c c h h h - + - + + + = .实际的控制权t D 应该是t<br>代人的子集的分配,以一种每代人都得到准确的决定相应集<br>2 .<br>第三个允许存在一定范围契约的选择是假设两代之间契约--我们假设t与t+1代人共同<br>以一代人中的每对都得到准确的决定相应集的方式被赋予对t t 1 D D+ 的子集的控制权, 然<br>后,在t 与t-1 代人的契约已知的情况下,他们选择一个可以使他们效用的加权平均数<br>t t 1 U lU + + 最大化的契约.注意一个特例,当l =0时两代人之间没有契约.<br>2 基准: 贝克瑞恩模型<br>这是由标准偏好决定的模型,即不存在象征性消费,与完全信用市场相连的完全利他主<br>义,以及不存在契约.前两个假设隐含着给定每代人的偏好为1.完全信用市场提供任何人都<br>可以按市场利率r来开展他们想要的任何数目的借贷的假设.在这种假设下,每个家族都存<br>在现实中的预算约束.<br>其中t w 是第t代人的原始财富. ( ( ) ) ( ) ( ) 1 1 H 1 L , 1<br>t t t t g w g w s h h e + + - - 项代表以商品单<br>位计算的对下代人人力资本的投资.<br>因为不存在契约, 每代人当处于中年( ) t c 和老年( ) t 1 p + 时都被赋予对自己消费的控制<br>权以及教育自己孩子的权利( ) t 1 h + ,在1中已经给出的最大化效用方程在2条件下告诉我们:<br>注意到把当期的资助ht 和将来期望的水平t 2 h + 看成已知时, t 1 h + 完全由最后一个方程决<br>定.值得注意的是在方程中没有与效用有关的项.这里有两个隐含条件:第一,因为t 1 h + 的选择<br>_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _<br>2<br>换句话说<br>,<br>如果<br>t<br>代人决定向量{ , , } t t t t X = p c h (<br>大约<br>) ,<br>那么<br>t + s<br>代人将决定向量t s X +<br>.<br>不受消费边际效用高低的影响,所以不存在收入效应;第二,家长偏好不影响对人力资本投资<br>的决定. 这些结论都来自于这样的事实:在完全资本市场下,投资人力资本的决定可以完全用<br>他的收入净现值分析--今天对儿子进行教育投资的人可以很快以隐含着的他儿子将来收入<br>的增长将钱赢回,并进行消费. 3<br>评论1:在没有象征性消费,完全利他主义,不存在契约和完全信用市场的模型下,对人力<br>资本的投资既不依赖于家长有多少钱,也不依赖于他们的偏好.<br>3 Beckerian 模型之外<br>3.1 信用约束的作用<br>让我们现在想象一个没有信用市场，或者至少存在的信用市场对大多数人来说是不相关的<br>世界。除了已经在第二部分加的限制，这次我们对偏好等级不多加限制。在这个世界，人们<br>用家庭资源来投资孩子的学校教育。此外，投资教育是家庭唯一可及的投资机会。但是，和<br>前面的章节一样，这里也有税和补贴。所以，任何时期的预算约束都可以写成：<br>对每一个t，这种约束都成立。此外，为了使信用约束有意义，消费总是被设定为非负。这<br>个假设排除了投资人力资本的决定可以用对收入现值的影响来衡量的可能性--投资水平的<br>改变必定会引起一定时期和/或人们之间消费分配的改变.因此,消费边际效用必须加入决定<br>教育投资的计算中去,而且,正因为收入的改变对边际消费效用有影响,它们将影响对人力资<br>本的投资.更特别地,从公式6可以很明显的看到当期收入的任何增加( o e 或<br>L<br>t w 的增加)将导致<br>t 1 h + , t c 或t p 的增加.假设t 1 h + 没有增加,那么至少t c , t p 中的一个必须要增加,与之相关的边际<br>效用也必须增加.但是,因为我们的独立性假设, 对人力资本投资的额外所得不变和由此导<br>致-------------------------------------------------------------------------------------------------------------------<br>---<br>3但是,虽然在这个经济中不存在收入效应,可能会出现一个在数据中看起来很像财富效应的父母人力资本<br>效应<br>:<br>例如<br>,<br>如果12 s p 0,<br>在t 2 h + 不变的情况下<br>,<br>t h 的增长将使教育投资的边际成本降低<br>(<br>不影响收益<br>),<br>因此<br>导致t 1 h + 的提高<br>.<br>事实上<br>,<br>所有代人人力资本的数目都将提高<br>.<br>t 1 h + 的增长导致t 2 h + 的增长<br>,<br>从而使<br>2 2 1 ( , ) t t s h h + + 下降<br>,<br>这又引起t 1 h + 的进一步上升<br>,<br>等等<br>.<br>的与相关的新收入t 1 h + 不变以及t c , t p 的增长并不是一个最佳选择.所以,必须增加t 1 h + .收入<br>收入的增长必须与人力资本投资的增长有关.此外,鉴于集体偏好中的多种收入,它能够达到<br>的增长程度很明显与相关的权数有关,因此通常与父母偏好有关.<br>这个逻辑的成立是比较容易的，但是在这里被允许的广泛偏好使这个逻辑变的很麻烦。<br>所以，我们把自身限定在不正式的陈述隐含的结果。<br>评论二：只要信用市场是不完全的，收入水平和/或者偏好不同的两个家庭则通常会投资<br>不同数量的人力资本。即使完全是利他主义，没有象征性消费，评论仍然成立。<br>3.2 非标准偏好的角色<br>在这一部分,我们再次使用完全信用市场的假设,但现在扩展一系列偏好以允许象征性消<br>费和不完全利他主义.但并不完全这么简单,特别是即便对贝克瑞恩公式一个很小的偏离都会<br>导致意想不到的复杂后果.为了看到Barro-Becker 偏好假设(即t 代人的效用方程<br>0 ( 1 ) s M ,<br>s t s t s d U c p = + + + )与两代间契约假设(取代无契约)的结合,令t与t+1代人具有对t c ,<br>t 1 p + 和t 1 h + 的控制权. 那么t 与t+1 代人将希望<br>( ) 1 ( )<br>0 1 1 1 s M , s M ,<br>s t s t s s t s t s d U c p l d -U c p<br>= + + + = + + + 邋+ 最大化,这给他们提供了一个有效的<br>效用方程<br>这个效用函数是用双曲线贴现的形式建立起来的.文学作品(Harris-Laibson(2000))中一个广<br>为人知的事实是这导致持续时间不长的决定规则,因此当代人在做决定时必须考虑到下一批<br>决策者的反应函数. 贝克瑞恩模型通过做出l =0的特定假设来避免了这个问题. 我们现在来<br>考虑一个存在教育支出的象征性消费与潜在非完全利他主义的模型.<br>这里隐含一个没有税收与补贴的假设.为了完善对这个模型的描述,假设t与t+1之间的两代<br>契约对t c , t 1 p + 和t 1 h + 有控制权.前面已经提到, 把这个案例归属于没有契约当中去. 最大化需<br>要服从预算约束(方程6)的事实告诉我们在内在最大化条件下它必须是这种情况:<br>注意:在写下这种情况时我们必须处理时间连贯性问题.这是因为我们同时假设完全信用市场<br>以及s12 = 0 .如果12 s 不为零, 将来教育投资的边际成本将取决于当代人投资的多少,因此当<br>地人进行投资决策时必须要考虑到未来投资对当代投资的影响.<br>从方程7中可以看出,只有<br>UM 和<br>US 服从边际效用递减规律,家庭收入(即t h )的任何增<br>长都必定导致t 1 h + 的增长. 此外,由方程<br>UM 和<br>US 的性质得到的父母偏好很显然起了作用.<br>还有,很明显这儿的结果不受我们已假设的利他主义的性质与形式的影响,原因是这些项并没<br>有被考虑进以上的情况中.关键的条件是s 0<br>h U :当s 0<br>h U = 时,上面的情况减少至方程4,即我<br>们使之瓦解从而重新回到基准事例中.再次强调,不管是完全还是不完全利他主义以及我们是<br>不是没有契约(l = 0 )或两代人之间的契约,它都保持正确.另外,既然其他人物质消费的象征<br>性消费只是利他主义的一种有效形式,添加这种象征性消费并不改变这个结果.最后,尽管没<br>有在这儿说明, 很容易说明在包括完全契约在内的多代人契约下同样的结果(即除非存在人<br>力资本的象征性消费投资否则我们将得到与基准案例相同的结果)也成立.<br>这导致:<br>评论3: 在完全信用市场条件下,当且仅当存在人力资本投资的象征性消费时可能会有收<br>入效应和家庭偏好对人力资本投资的影响.<br>4 我们已经学过的<br>从这三个评论中我们可以看出当且仅当我们假设实质上不存在教育投资的象征性消费的情<br>况下,教育投资(参见Jacoby,1994;Glew与Jacoby,2000;以及Carvalho,2000)对收入效应的证据<br>可以被理解为信用约束的证据. 通过这些还使我们知道对于父母偏好也有类似的结果,如果<br>我们有对影响父母偏好的因素的好的计量办法,我们可以使用它们来研究信用约束,而方式与<br>我们使用收入效应完全一样。 <!--v:3.2--> ]]></description>
<category><![CDATA[政治思想]]></category>
<author><![CDATA[550115938@qq.com(wildyouth)]]></author>
<comments>http://550115938.qzone.qq.com/blog/1211424431#comment</comments>
<qz:effect>512</qz:effect>
<pubDate>Thu, 22 May 2008 02:47:11 GMT</pubDate>
<guid>http://550115938.qzone.qq.com/blog/1211424431</guid>
</item>

<item>
<title><![CDATA[Gravitational Lensing from Hamiltonian]]></title>
<link>http://550115938.qzone.qq.com/blog/1211424352</link>
<description><![CDATA[Massachusetts Institute of Technology<br>Department of Physics<br>Physics 8.962 Spring 1999<br>Gravitational Lensing from Hamiltonian<br>Dynamics<br><br>?1999 Edmund Bertschinger and Rennan Barkana.<br>Introduction<br>The deflection of light by massive bodies is an old problem having few pedagogical treatments.<br>The full machinery of general relativity seems like a sledge hammer when applied to weak gravitational fields. On the other hand, photons are relativistic particles and their propagation over cosmological distances demands more than Newtonian dynamics. In fact, for weak gravitational fields or for small perturbations of a simple cosmological<br>model, it is possible to discuss gravitational lensing in a weak-field limit similar to Newtonian dynamics, albeit with light being deflected twice as much by gravity as a nonrelativistic particle.<br>The most common formalism for deriving the equations of gravitational lensing is based on Fermat’s principle: light follows paths that minimize the time of arrival (Scheider<br>et al. 1992). As we will show, light is deflected by weak static gravitational fields as though it travels in a medium with variable index of refraction n =1 ? 2φ where φ is the dimensionless gravitational potential.<br>With the framework of Hamiltonian dynamics given in the notes Hamiltonian Dynamics<br>of Particle Motion, here we present a synopsis of the theory of gravitational lensing. The Hamiltonian formulation begins with general relativity and makes clear the approximations which are made at each step. It allows us to derive Fermat’s least time principle in a weak gravitational field and to calculate the relative time delay when lensing produces multiple images. It is easily applied to lensing in cosmology, including a correct treatment of the inhomogeneity along the line of sight, by taking advantage of the standard formalism for perturbed cosmological models.<br>Portions of these notes are based on a chapter in the PhD thesis of Barkana (1997). 1<br>Hamiltonian Dynamics of Light<br>Starting from the notes Hamiltonian Dynamics of Particle Motion (Bertschinger 1999), we recall that geodesic motion of a particle of mass m in a metric gμν is equivalent to Hamiltonian motion in 3 + 1 spacetime with Hamiltonian<br>0i<br>gpi<br>H(pi, xj , t) = ?p0 = gpi +(gij pipj + m2)+ ? 0i?2 ?1/2 . (1)<br>g00 ?g00 g00<br>2<br>This Hamiltonian is obtained by solving gμν pμpν = ?mfor p0. The spacetime coordinates<br>xμ =(t, xi) are arbitrary aside from the requirement that g00 &lt; 0 so that t is timelike and is therefore a good parameter for timelike and null curves. The canonical momenta are the spatial components of the 4-momentum one-form pμ. The inverse metric<br>components gμν are, in general, functions of xi and t. With this Hamiltonian, the exact spacetime geodesics are given by the solutions of Hamilton’s equations<br>dxi ?H dpi ?H<br>== . (2)<br>dt ?pi , dt ? ?xi<br>Our next step is to determine the Hamiltonian for the problem at hand, which requires specifying a metric. Because we haven’t yet derived the Einstein field equations, all we can do is to pick an ad hoc metric. In order to obtain useful results, we will choose a physical metric representing a realistic cosmological model, an expanding Big Bang cosmology (a Robertson-Walker spacetime) superposed with small-amplitude spacetime curvature fluctuations arising from spatial variations in the matter density. For now, the reader will have to accept the exact form of the metric without proof.<br>The line element for our metric is<br>2<br>ds2 = a (t) ?(1 + 2φ)dt2 + (1 ? 2φ)γij dxidxj . (3)<br>In the literature, t is called ?conformal? time and xi are ?comoving? spatial coordinates. The cosmic expansion scale factor is a(t) and is related to the redshift of light emitted at time t by a(t)=1/(1 + z). To get the non-cosmological limit (weak gravitational fields in Minkowski spacetime), one simply sets a = 1. The Newtonian gravitational potential φ(xi, t) obeys (to a good approximation) the Poisson equation. (In cosmology, the source for φ is not ρ but rather ρ ? ρˉwhere ˉ<br>ρ is the mean mass density; we will show this in more detail later in the course.) We assume |φ|? 1 which is consistent with cosmological observations implying φ ～ 10?5 .<br>In equation (3) we write γij (xk ) as the 3-metric of spatial hypersurfaces in the unperturbed<br>Robertson-Walker space. For a flat space (the most popular model with theorists, and consistent with observations to date), we could adopt Cartesian coordinates for 2<br>which γij = δij . However, to allow for easy generalization to nonflat spaces as well as non-Cartesian coordinates in flat space we shall leave γij unspecified for the moment.<br>Substituting the metric implied by equation (3) into equation (1) with m = 0 yields the Hamiltonian for a photon:<br>H(pi, xj , t) = p(1 + 2φ) ,p ≡ ? γij pipj ?1/2 . (4)<br>gWe have neglected all terms of higher order than linear in φ. Not surprisingly, in a perturbed spacetime the Hamiltonian equals the momentum plus a small correction for gravity. However, it differs from the proper energy measured by a stationary observer, E = ?V μpμ, because the 4-velocity of such an observer is V μ = (a(1 ? φ), 0, 0, 0) (since μν V μV ν = ?1) so that E = a?1p(1 + φ). The latter expression is easy to understand because a?1 converts comoving to proper energy (the cosmological redshift) and in the Newtonian limit φ is the gravitational energy per unit mass (energy).<br>Why is the Hamiltonian not equal to the energy? The answer is because it is conjugate to the time coordinate t which does not measure proper time. The job of the Hamiltonian is to provide the equations of motion and not to equal the energy. The factor of 2 in equation (4) is important ? it is responsible for the fact that light is deflected twice as much as nonrelativistic particles in a gravitational field.<br>To first order in φ, Hamilton’s equations applied to equation (4) yield<br>dxiidpi<br>i<br>dt = n (1 + 2φ) , dt = ?2p?iφ + γkij pk nj (1 + 2φ) ,n ≡ γij ppj . (5)<br>We will drop terms O(φ2) throughout. We have defined a unit three-vector ni in the photon’s direction of motion (normalized so that γij ninj = 1). The symbol γkij = 1 γkl(?iγjl + ?j γil ? ?lγij ) is a connection coefficient for the spatial metric that vanishes<br>2<br>if we are in flat space and use Cartesian coordinates. Beware that ?i is the covariant derivative with respect to the 3-metric γij and not the covariant derivative with respect to γμν , although there is no difference for a spatial scalar field: ?iφ = ?iφ.<br>Note that the cosmological expansion factor has dropped out of equations (5). These equations are identical to what would be obtained for the deflection of light in a perturbed Minkowski spacetime. The reason for this is that the metric of equation (3) differs from the non-cosmological one solely by the factor a2(t) multiplying every term. This is called a conformal factor because it leaves angles invariant. In particular, it leaves null cones invariant, and therefore is absent from the equations of motion for massless particles.<br>In the following sections we shall represent three-vectors (and two-vectors) in the 3-space with metric γij using arrows above the symbol. To lowest order in φ, we may interpret these formulae as giving the deflection of light in an unperturbed spacetime due to gravitational forces, just as in Newtonian mechanics. The difference is that our results are fully consistent with general relativity. 3<br>?<br>? ?<br>?<br>3<br>Fermat’s Principle<br>When ?tφ = 0, the Hamiltonian (eq. 4) is conserved along phase space trajectories and the equations of motion follow from an alternative variational principle, Maupertuis’ principle (Bertschinger 1999). Maupertuis’ principle states that if ?H(pi, qj , t)/?t = 0, then the solution trajectories of the full Hamiltonian evolution are given by extrema of<br>the reduced action pi dqi with fixed endpoints. This occurs because<br>pi dqi ? H dt = pi dqi ? d(Ht) + t dH . (6)<br>The Ht term, being a total derivative, vanishes for variations with fixed endpoints. The t dH term vanishes for trajectories that satisfy energy conservation, and we already know (from the Hamilton’s equations of the full action) that only such trajectories need be<br>considered when ?H/?t = 0. Thus, the condition δpi dqi = 0, when supplemented by conservation of H, is equivalent to the original action principle.<br>Expressing pi in terms of dxi/dt using Hamilton’s equations (5) in the full phase space for the Hamiltonian of equation (4), the reduced action becomes<br>pi dxi = pγij nj dxi = H(1 ? 2φ)γij n i dxj = H dt . (7)<br>Using H = constant ≡ h, Mauptertuis’ principle yields Fermat’s principle of least time,<br>?? ??1/2<br>dxi dxj<br>δ dt = δ [1 ? 2φ(x)] γij ds = 0 (8)<br>ds ds<br>for light paths parameterized by s. We leave it as an exercise for the reader to show, using the Euler-Lagrange equations, that if s measures path length, equation (8) yields equations (5) exactly (to lowest order in φ) when ?tφ = 0. In comparing with equation (5), one must be careful to note that there the trajectory is parameterized by dt = (1 ? 2φ)ds so that ?n = d?x/ds is a unit vector.<br>Thus, for a static potential φ (even in a non-static cosmological model with expansion factor a(t)), light travels along paths that minimize travel time but not path length (as measured by the spatial metric γij ). The null geodesics behave as though traveling through a medium with index of refraction 1 ? 2φ. To minimize travel time, light rays will tend to avoid regions of negative φ; therefore light will be deflected around massive bodies.<br>Fermat’s principle is exact for gravitational lensing only with static potentials. In most astrophysical applications, the potentials are sufficiently relaxed so that ?tφ may be neglected relative to ni?iφ and Fermat’s principle still applies. The one notable exception<br>is microlensing, where the lensing is caused by stars (or other condensed objects) moving across the line of sight. In this case, one may still apply Fermat’s principle after boosting to the rest frame of the lens. 4<br>4<br>Reduction to the Image Plane<br>In equation (8), the action is invariant under an arbitrary change of parameter, ss?(s)<br>→<br>with ds?/ds &gt; 0. This is not a physical symmetry of the dynamics, and as a consequence we may eliminate a degree of freedom by using one of the coordinates to parameterize the trajectories. A similar procedure was used to eliminate t in going from equation (6) to equation (8). Here, as there, the Lagrangian is independent of the time parameter, enabling a reduction of order. However, for reasons that will soon become clear, this reduction cannot be done using the reduced action (Maupertuis’ principle) but instead follows from reparameterization of the Lagrangian.<br>To clarify the steps, we start with<br>? dxi dxj ?1/2<br>i<br>L3(x , dxj /ds) = [1 ? 2φ(x)] γij (9)<br>ds ds<br>for the Lagrangian in the three-dimensional configuration space (eq. 8). Because the Lagrangian does not depend explicitly on s, the Hamiltonian is conserved and we may attempt to reduce the order as in the previous section. The first step is to construct the Hamiltonian. Under a Legendre transformation, L3 → H3(pi, xj , s)= pi(dxi/ds) ? L3<br>i<br>where pi = ?L3/?(dxi/ds) is the momentum conjugate to x. But we quickly run into trouble: as the reader may easily show, H3 vanishes identically.<br>What causes this horror? The answer is that L3 is homogeneous of first degree in the coordinate velocity dxi/ds, which is equivalent to the statement that the action of equation (8) is invariant under reparameterization. Physically, the Hamiltonian vanishes because of the extra symmetry of the Lagrangian, which is unrelated to the dynamics. The physical Hamiltonian should include only the physical degrees of freedom, so we must eliminate the reparameterization-invariance if we are to use Hamiltonian methods.<br>This is done very simply by rewriting the action (eq. 8) using one of the coordinates as the parameter. The radial distance from the observer is a good choice: for small deflections of rays traveling nearly in the radial direction toward the observer, r will be single-valued along a trajectory.<br>To fix the parameterization we must write the spatial line element in a Robertson-Walker space in terms of r and two angular coordinates:<br>dl2 ≡ γij dxidxj = dr2 + R2(r)γab(ξ)dξadξb . (10)<br>Here 1 ≤ a, b ≤ 2 and γab is the metric of a unit 2-sphere. The coordinates ξa are angles and are dimensionless. Note that r measures radial distance (γrr = 1) and R(r) measures angular distance. We will not give the exact form of R(r) here except to note that for a flat space, R(r) = r. In the standard spherical coordinates, γθθ = 1 and γφφ = sin2 θ. We will leave the coordinates in the sphere arbitrary for the moment, and use γab and its inverse γab to lower and raise indices of two-vectors and one-forms in the sphere. 5<br>Our action, equation (8), is the total elapsed light-travel time t (using our original spacetime coordinates, eq. 3). The reparameterization means that now we express the action as a functional of the two-dimensional trajectory ξa(r):<br>? rS ??1/2<br>dxa dxb<br>t[ξa(r)] = [1 ? 2φ(ξ, r)] 1 + R2(r)γab dr . (11)<br>dr dr<br>0<br>This action is to be varied subject δξa = 0 at r = 0 (the observer) and r = rS (the source).<br>In writing equation (11), we have neglected ?φ/?t and we have neglected terms O(φ2) (weak-field approximation). As we will see, the angular term inside the Lagrangian is small when the potential is small, and therefore we can expand the square root, dropping all but the lowest-order terms. To the same order of approximation, we may neglect the curvature of the unit sphere, and set γab = δab. (We can always orient spherical coordinates so that γab = δab plus second-order corrections in ξ.) These approximations<br>together constitute the small-angle approximation. In practice it is well satisfied; observed angular deflections of astrophysical lenses are much less than 10?3 .<br>With the weak-field and small-angle approximations, the action becomes<br>? rS ??<br>dξb 1 dξa dξb<br>t[ξa(r)] = rS + L2 dr , L2 ξa , , r = R2(r)δab dr dr ? 2φ(ξa , r) . (12)<br>dr 2<br>0<br>Note that the Lagrangian now depends on the ?time? parameter, so we have eliminated the parameterization-invariance.<br>To get a Hamiltonian system, we make the Legendre transformation of the Lagrangian L2. The conjugate momentum is pa = R2(r)δabdξb/dr. The Hamiltonian becomes<br>2<br>H(pa, ξb , r) = 2Rp2(r) + 2φ(?<br>ξ, r) . (13)<br>On account of the small-angle approximation, ?<br>≡ δab<br>p and ξ?are two-dimensional vectors in Euclidean space (p2 papb). Noting that r plays the role of time, this Hamiltonian represents two-dimensional motion with a time-varying mass R2(r) and a time-dependent potential 2φ.<br>With the Hamiltonian of equation (13), Hamilton’s equations give<br>dξ?p p<br>?d??φ<br>dr = R2(r) , dr = ?2 ?ξ?. (14)<br>These equations and the action may be integrated subject to the ?initial? conditions ξ = ξ0, ?<br>p = 0 and t = t0 at the observer, r = 0<br>ξ?(r)= ξ?0 ? R(r) R(r ? r?) ?φ (ξ?(r?), r?) dr?<br>R(r?) ?ξ?<br>0<br>? r<br>?φ<br>p(r)= ?2(ξ?(r?), r?) dr? (15) 0 ?ξ?<br>? r ? p2(r?)<br>t(r)= t0 ? r ? R2(r) ? 2φ(ξ?(r?), r?) dr? .<br>0 ?<br>Note that here t is the coordinate time along the past light cone; the elapsed time (the action) is t0 ? t. The two terms in the time delay integral arise from geometric path length (the p2 term) and gravity. Half of the gravitational potential part comes from the slowing down of clocks in a gravitational field (gravitational redshift) and the other half comes from the extra proper distance caused by the gravitational distortion of space.<br>Equations (15) provide only a formal solution, since φ is evaluated on the unknown path ξ?(r?). The reader may verify the solution by inserting into equations (14). One needs the following identity for the angular distance in a Robertson-Walker space, which we present without proof: ?<br>?R(r ? r?)= 1 . (16)?r R(r)R(r?) R2(r)<br>It is easy to verify this for the flat case R(r) = r.<br>When the potential varies with time, we cannot use Fermat’s principle or the further reduction achieved in this section. Instead, one has to integrate the original equations of motion (5). It can be shown (Barkana 1997) that, under the small-angle approximation, these equations also have the formal solution given by equation (15), with the single change that φ also becomes a function of t and that t must be evaluated along the trajectory: φ(ξ?(r?),r?, t(r?)). Thus, we obtain the physical result that the potential is to be evaluated along the backward light cone.<br>Astrophysical Gravitational Lensing<br>The astrophysical application of gravitational lensing is based on the following considerations.<br><br> <!--v:3.2--> ]]></description>
<category><![CDATA[自然哲学]]></category>
<author><![CDATA[550115938@qq.com(wildyouth)]]></author>
<comments>http://550115938.qzone.qq.com/blog/1211424352#comment</comments>
<qz:effect>512</qz:effect>
<pubDate>Thu, 22 May 2008 02:45:52 GMT</pubDate>
<guid>http://550115938.qzone.qq.com/blog/1211424352</guid>
</item>

<item>
<title><![CDATA[Externalities, the Coase Theorem and Market Remedies]]></title>
<link>http://550115938.qzone.qq.com/blog/1211424023</link>
<description><![CDATA[Externalities, the Coase Theorem and Market Remedies <br>David Autor <br>14.03 Spring 2004 <br>1 Introduction <br>  You were introduced to the topic of externalities in 14.01. An externality arises when an economic <br>actor does not face the ?correct price? for taking a speci?c action. The correct price of an <br>action is the marginal social cost of that action. As we discussed during the section on General <br>Equilibrium, when markets work properly, they align private costs and bene?ts with social <br>costs and bene?ts. When private bene?ts di¤er from social bene?ts (either higher or lower), <br>externalities result. <br>  Some classic externalities include: <br>?Tra￠ c: When I take the highway, I increase congestion for other drivers, a negative exter- <br>nality. Since the toll on the Mass Turnpike does not vary with tra￠ c conditions, I probably <br>face the wrong price of driving on the highway (too high at non-peak hours, too low at peak <br>hours). As a result, I use the Pike ?too much?during peak hours and not enough during <br>non-peak hours. <br>?Disease transmission: When I decide whether to have my children receive ?u shots, I con- <br>sider whether the cost of the inoculation in time, money, discomfort is worth the reduced <br>risk. I probably do not consider that by protecting my children from the ?u, I also protect <br>the children at their school. Because my private bene?t of the shot does not incorporate <br>the external social bene?t of the shot, I am less motivated than I ?should be? to get my <br>children inoculated. It is therefore likely that too few children receive small pox vaccines. <br>Ironically, there are other parents who recognize that, because most parents do get their <br>children inoculated, other kids may be reasonably protected even without receiving an <br>inoculation. Hence, these parents free-ride on the positive externality, and are even less <br>1 <br>motivated to get a shot than the parents who do not consider the positive externality they <br>are generating. This exacerbates the problem. <br>  Pollution: Because clean air is not priced, I pay essentially no cost to pollute the air. When I <br>decide whether to drive to work or take the train, my marginal cost of driving (fuel plus wear and <br>tear on my car) probably does not incorporate the social cost of additional pollution. Because <br>my private cost is lower than the social cost, I will likely drive ?too much?relative to a case where <br>I faced the full marginal social cost. <br>Are these externalities never ?internalized?by the market? <br>2 The Coase Theorem <br>  Until the publication of Ronald Coase?1960 paper, ?The Problem of Social Cost,?most econo- <br>mists would have answered yes. Coase made them reconsider that view. <br>  Coase gave the example of a doctor and a baker who share an o￠ ce building. The problem: <br>the baker?s loud machinery disturbed the doctor?s medical practice. The doctor could not treat <br>patients while the baker?s machinery was running. <br>  The standard economic reasoning (at the time) was that the baker should have to compensate <br>the doctor for the harm he was causing since he was ?causing?the externality. Having the baker <br>provide compensation would correct the externality imposed on the doctor. <br>  But is this reasoning complete? Coase pointed out that one could re-frame this problem as <br>follows: a doctor sets up his o￠ ce in a new building and after moving in notices that the baker?s <br>machinery is too loud for him to conduct his practice. He demands that the baker shut down <br>his operation due to the disturbance. <br>  Who is responsible for the externality in this case? One can legitimately argue that the doctor <br>is creating an externality by requiring the baker to bake in silence. The baker?s noise can be <br>viewed as an ?input?into his production of baked goods. Without it, the baker could not perform <br>his work. So perhaps the doctor should accommodate the baker, either by moving his practice <br>or by installing soundproo?ng. <br>  If this reasoning is valid, then who should compensate whom? From a legal point of view, the <br>answer may be clear. From an economic point of view, the answer is indeterminate based only <br>on the information provided. <br>2 <br>  Consider the following additional information. The baker could buy quieter machinery for $50. <br>The Doctor could soundproof his walls for $100: Economic e￠ ciency demands that the lowest <br>(marginal) cost solution that achieves the objective is the right solution. The baker should buy <br>quieter machinery. <br>  So, does this mean that the baker should have to pay to abate? It does not. <br>  Consider the following scenarios: <br>1. The town council assigns the doctor the right to control the noise level in the building. He <br>noti?es the baker that he needs quiet to work. The baker spends $50 for quieter machinery. <br>2. The town council assigns the baker the right to make as much noise as needed to do his <br>work. The doctor complains about the noise and the baker points out that he has the right <br>to make as much noise as he likes. Will the doctor now spend $100 to sound proof his <br>o￠ ce? If the doctor and baker can negotiate readily, they should arrange for the doctor to <br>pay the baker $50 to buy quieter machinery. <br>  As this example demonstrates, the e￠ cient economic outcome should occur regardless of which <br>party is ?responsible?for the externality. In either case, quieter baking machinery is purchased. <br>  However, the legal framework does matter. If the ?sound rights?are assigned to the doctor, the <br>baker spends loses $50. If the sound rights are assigned to the baker, the doctor spends $50: <br>  So the Coase theorem says the following. If (1) property rights are complete (so, in our example, <br>one party clearly owns the ?sound rights?) and (2) parties can negotiate costlessly (so the doctor <br>and baker don?t come to blows), then the parties will always negotiate an e￠ cient solution to <br>the externality. The law determines who pays this cost, but the outcome is the same. (Note the <br>parallelism with the Welfare theorems: e￠ ciency and distribution are separable problems.) <br>  The Coase theorem implies that the market will solve externalities all by itself unless: (1) <br>property rights are incomplete (for example, no one owns the air) or (2) negotiating is costly <br>(for example, the entire population owns the air, but all citizens cannot simultaneously negotiate <br>about pollution levels). <br>  The Coase theorem is often misinterpreted to suggest that the market will solve all externalities. <br>This is not true, and Coase will probably go to his grave railing against the ?Coaseans?who make <br>this claim. <br>3 <br>  Rather, the Coase theorem suggests that the market can potentially solve externalities if prop- <br>erty rights are clearly assigned and negotiation is feasible. <br>  In some cases, this is clearly infeasible. <br>?Airlines cannot realistically negotiate with individual homeowners for over?ight rights to <br>their houses, even though these over?ights do create externalities. <br>?I cannot negotiate with all handicapped drivers for the use of an empty handicap space in <br>an emergency, even though I?d be glad to pay these drivers handsomely to rent the parking <br>space. <br>  A key inference that follows from the Coase Theorem is that the best solution to resolving an <br>externality may not be to regulate the externality out of existence but rather to assign property <br>rights or facilitate bargaining so that the relevant parties can ?nd an economically e￠ cient <br>solution. <br>3 Remedying pollution: Three approaches <br>  Consider two oil re?neries that both produce fuel, which has a market price of $3 per gallon <br>(assume demand is in?nitely elastic so that this price is ?xed regardless of the quantity produced). <br>  Assume that each re?nery uses $2 in raw inputs (crude oil, electricity, labor) to produce 1 gallon <br>of fuel. <br>  In addition, each plant produces smog, which creates $0:01 of environmental damage per cubic <br>foot. <br>  The amount of smog per gallon of fuel produced di¤ers at the two plants: <br>s1 = y2 <br>1; <br>s2 = <br>1 <br>2 <br>y2 <br>2; <br>where y1; y2 denote the number of gallons of fuel produced at each plant. So, plant 2 pollutes <br>only 1 <br>2 as much for given production. <br>  Assuming initially that there are no pollution laws. In this case, each plant will produce as many <br>gallons as it can until it runs out of capacity (since it makes $1 pro?t per gallon). Assume each <br>plant can produce 200 gallons. <br>4 <br>3.1 Competitive outcome <br>  What will ?rms optimally do in the absence of any legal framework for resolving the externality? <br>The problems for the respective ?rms are: <br>max <br>y1 <br>1 = y1   (3 ?? 2) s:t: y1   200; <br>max <br>y2 <br>2 = y2   (3 ?? 2) s:t: y2   200; <br>y 1 = y 2 = 200: <br>  Each ?rm ignores the social damage from its smog production (notice that s1; s2 do not enter <br>into the ?rms?pro?t maximization problems). Hence, pollution is s1 = 40; 000; s2 = 20; 000. <br>The negative pollution externality is $400 and 200 from plants 1 and 2 respectively. <br>3.2 Welfare maximizing case <br>  Before analyzing how to correct this externality, we need to determine the ?optimal? level of <br>pollution. In this case, optimal pollution is non-zero. More generally, not all activities that <br>generate externalities should be stopped. But if these activities generate negative (positive) <br>externalities, then social e￠ ciency generally suggests that we want to do less (more) of them <br>than would occur in the free market equilibrium. <br>  To get the socially e￠ cient level of fuel production, we want to equate the marginal social bene?t <br>of the last gallon of fuel to the marginal social cost. <br>  What is the social bene?t? It is $3. This comes from the in?nitely elastic demand curve. <br>  The marginal social cost of production is $2 in raw inputs plus whatever pollution is produced. <br>  The e￠ ciency condition is MBs = MCs; marginal social bene?t equals marginal social cost. <br>  We therefore want it to be the case that at the margin, that no more than $1 of environmental <br>damage is done per gallon of fuel produced. Consequently, no plant should produce more than <br>100 cubic feet of smog per gallon of fuel. <br>  (Note: no plant should produce less than this amount either since the pollution is indirectly <br>bene?cial: it is an ?input?into the production process; less pollution means less fuel production). <br>5 <br>  Imagine that each plant faced the private plus social costs of production. If so, we could rewrite <br>the previous pro?t maximizing conditions as: <br>max <br>y1 <br>1 = y1   (3 ?? 2) ?? 0:01y2 <br>1 s:t: y1   200; <br>max <br>y2 <br>2 = y2   (3 ?? 2) ?? 0:005y2 <br>2 s:t: y2   200; <br>y   1 = 50; y 2 = 100: <br>  When Plant 1 is producing 50 gallons, the marginal gallon produces 100 cubic feet of smog, which <br>causes $1:00 in environmental damage. More pollution than this would be socially ine￠ cient <br>since fuel sells for $3 and uses $2 in raw inputs to produce. For Plant 2, the corresponding <br>production is 100 gallons because this plant produces less smog per gallon. <br>We now have an e￠ cient benchmark for welfare maximization. <br>How do we get plants to produce the socially e￠ cient level of pollution? Three classes of regulatory <br>solution are possible. Each has di¤erent properties. <br>3.3 Command and control (?quantity?) regulation <br>  ?Command and control? regulation is the traditional approach to limiting externalities. This <br>approach sets numerical quantity limits on activities that have external e¤ects. It is often called <br>?quantity?regulation. <br>  The most common command and control regulation is simply banning an activity ??though <br>shalt not litter.? But as we know, just because an activity has external e¤ect does not mean it <br>should be banned outright?only that too much or too little relative may be done relative to the <br>social optimum. <br>  Much command and control regulation recognizes this point, and so permits some positive <br>amount of an activity, but less than a private actor would otherwise undertake. <br>  How does this apply to the example above? We know the optimal quantity of production for <br>each plant from our calculations above. We could therefore pass a law that says ?Plant 1 may <br>produce 50 gallons of fuel and Plant 2 may produce 100 gallons of fuel.?This will achieve exactly <br>the desired result. <br>  But this kind of regulation is clumsy. <br>?It?s di￠ cult to write laws that regulate the behavior of each plant individually. <br>6 <br>?Once passed, such laws are di￠ cult to modify as technology or pollution costs change. <br>?If the law cannot be written to regulate each plant?s output di¤erentially, further ine￠ cien- <br>cies will result. <br>?[For example, calculate as an exercise the optimal amount of fuel production to permit these <br>two plants to produce if the regulator must apply the same numerical production cap (in <br>fuel) for each plant. (Hint: the answer is not 75 gallons.) This is actually a commonplace <br>case: regulators can set average output at the industry level but cannot further regulate <br>the behavior of individual plants. As your calculation shows, this leads to ine￠ ciencies <br>where some regulated plants pollute ?too much? and other regulated plants pollute ?too <br>little?relative to the e￠ ciency condition that MBs = MCs.] <br>  Despite these weaknesses, command and control regulation is the most common approach taken <br>for regulating externalities. <br>3.4 Pigouvian tax (?price regulation?) <br>  An alternative approach is to use the price system to ?internalize?the externality. <br>  We know from above that the marginal social cost of pollution is $0:01 per cubic foot of smog. <br>If we charged ?rms for polluting, the social cost would be incorporated in the private cost. Done <br>correctly, ?rms will make optimal choices. <br>  This type of tax is known as a Pigouvian tax after the economist Pigou who suggested it. <br>  Speci?cally, it we set the pollution tax at t = $0:01 per cubic foot of smog, then each plant <br>would choose the optimal quantities due to its pro?t maximization. In other words <br>max <br>y1 <br>  = y1(3 ?? 2) ?? t   y2 <br>1; where t = 0:01 ! yp <br>1 = 50 <br>max <br>y2 <br>  = y2(3 ?? 2) ?? t   <br>1 <br>2 <br>y2 <br>1; where t = 0:01 ! yp <br>2 = 100 <br>  This solution achieves the desired result with arguably less complexity. Facing this tax, plants <br>will choose the e￠ cient amount of production. We do not have to write a separate law for each <br>plant. <br><br><br> <!--v:3.2--> ]]></description>
<category><![CDATA[自然哲学]]></category>
<author><![CDATA[550115938@qq.com(wildyouth)]]></author>
<comments>http://550115938.qzone.qq.com/blog/1211424023#comment</comments>
<qz:effect>512</qz:effect>
<pubDate>Thu, 22 May 2008 02:40:23 GMT</pubDate>
<guid>http://550115938.qzone.qq.com/blog/1211424023</guid>
</item>

<item>
<title><![CDATA[物理的挑战3—李政道]]></title>
<link>http://550115938.qzone.qq.com/blog/1210483357</link>
<description><![CDATA[<span style="font-size:16px;line-height:1.8em;">李政道：</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    近代的物理，它有三个相当重要的对称，一个是左右对称，就是通常也可以叫宇称守恒跟不守恒，一个是正反粒子，正粒子变成反粒子，一个是时间反演，过去、将来的对称，这三个是跟自然界，整个宇宙的演变有极为极为密切的关系，跟我们的存在有极为密切的关系。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    我第一步就先讲左右对称，当然大家都知道左并不是右，左跟右当然不一样，可是左右是不是对称这句话是什么意思，你看左右不一样，它的字就不一样的，它的字就不一样，左是这样写的，右是这样写的，在几年以前我去西安博物馆，看到它的竹简文，它竹简文里面这个右（字），不是像我们这样写这个右（字），它是左（字）的反过来，它的右（字）是这样一个右（字），所以，我看了很有启发，启发呢，我就写了四句，说汉代右（字），这是汉朝时候的右（字），右（字）以镜中左（字），在汉朝这个右字是镜子里面的左字，今日反而写为右（字），这是现在的右（字），所以左右两字不对称，汉朝（时）是对称的，宇称守恒也不准，所以汉朝的时候，左右是对称的。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    我们现在是不对称了，不对称是什么意思，就是开头的时候，左右完全一样，以后除了左右都不一样，而且我们可以算出来怎么不一样，另外一个有关的，就是π介子，嬗变以后，这也是1957年，加温跟莱德曼他们做的，它嬗变到电子，负电子跟正电子，不光是左右不一样，他也发现正粒子跟反粒子，你不变左右，就把正粒子反粒子一变，也不一样，这样的我们叫正粒子反粒子不对称，我们叫C不对称，再进一步还有一个不对称，它里面有三道不对称，一个是左右，另一个是正粒子反粒子（不对称），第三个（是）时间，将来跟过去（不对称），这三个是联起来的。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    这个重要的实验是1964年做的，所以我们先讲讲正粒子和反粒子，普通的概念是电子的电荷是负的，也是我们的问题，电子的电荷为什么是负的呢？人家说这样，因为质子也有电，它的电子电荷是正的，因为电荷相反的相吸，所以负电子跟质子的正电荷相吸，所以成为一个原子，既然质子的电荷是正的，（所以）电子的电荷就是负的，那么你可以说，质子的电荷为什么是正的呢？到底是因为电子的电荷是负的，它要相吸，就是说本来在1957年以前或者1956年以前，电子是正的跟负的，正电子当然不是负电子，哪个是正电子，哪个是负电子，是相对的一个（概念），就跟左右一样，左不是右，可它是对称的，这个呢，我们发现并不是，不光是把正电子变成负电子，正电子变成负电子，左变右，它也不对称，这是一个非常重要的。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    这个实验是1964年做的，Crmide（和）Fich两个人都得诺贝尔奖了，他是把我们叫做长寿的K，K，你不需要太深了解，它是个圆圆的一个粒子，它没有电荷的，它有个寿命，不光是它没有电荷，（在）它任何一点，它没有任何的电流，这个你可以证明出来，所以完全电性是中性的，这样一个它衰变以后呢，它可以衰变到电子，也可以衰变到正电子地也可以（衰变）到负电子，两个都有的，但是知道它本身是没有电的这个思想的，你觉得它应该到正电子的数目到负电子的数目应该是一样，因为它本身是没有电流的思想在里面，这个量出来，不然，它到正电子，数目字比（到）负电子（数目字）多了千分之六，多了一点，就是说在这个我们叫K，就是中性的长寿的K介子，完全跟电子没有一点关系的，它死的时候，它的下一代，负的跟正的不一样，这个规律是相当惊人的，那么这个规律是在1964年这个规律我们就叫，它不光是C，C就是粒子跟反粒子，我们叫CP，正粒子，反粒子，左变右，还不一样，叫C  P  不对称，这个C  P不对称，我们等会儿底下的一个重点，CP不对称，你看上去，牵扯到时间，将来跟过去不对称，然后呢，把左变右，粒子变成反粒子，将来变成过去，又对称起来，这个整个宇宙的规律是这样，我再重复一下，因为底下讲这个是重点，这个重点就是宇宙开始的大爆炸到现在，我们为什么存在，我们在世界上最不懂的是哪几个（问题），而且很重要，这个一看就知道很重要，就是说我们的宇宙里面90％以上的物质不是我们的物质，我们宇宙里面有类星体，它的能量是远远远远超过了太阳能，这些都是跟宇宙的发展有关系的，都是跟P的不对称，C不对称，CP不对称，T不对称，跟CPT有密切关系，怎么样联系呢，我们还没有了解，这是现在研究的一个很大的一个方向。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    所以我再重复一下，它有三个基本的对称是很重要的，一个叫P，P呢Parity宇称，就是左右，什么叫左右不对称，左右当然不一样，左右不对称是什么意思？有两个组织，从开始时完全一样，就是左跟右不一样，在发展是，不光左不一样，其他也不一样，而且这个不一样我们可以操纵的，正负电子部对称，什么意思，正电子负电子当然不对称了，不过呢，你可以想，一个正电子跟负电子其他就是正负符号不一样，假如我们叫我们可以想象有一个物质是反物质，是正电子代替了我们现在的负电子，这是反，这是（正），那么它可以做个反氢，也可以做个世界了，不然它是不一样，这个我们叫C，就把正粒子变成反粒子，它不对称，不对称的意思是不光是这个不一样，以后发展其他也不一样，CP，CP把正粒子变成反粒子，左变右，也不一样，叫CP不守恒，T把时间，将来跟过去大家可以说，将来跟过去当然不一样了，有什么稀奇，这个我来，等会儿解释，可是，假如你把左变右，右变左，正粒子变成反粒子，将来变成过去，过去变成将来，又对称起来了，这个准确度是相当惊人的。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    二十世纪的物理发展是简化归纳，因为我们怎么会叫简化归纳二十世纪，我们说大（物质）是小的（物质）做的，小的（物质）是更好的（物质做的），我把最小的找着，我就了解最大的，这个很简单，这个方向从汤姆逊在十九世纪末年，发现电子，卢瑟福发现核子，这个里面是大发展，一直到现在，我们有很精密的仪器，科学设备，有精密的理论，相对论，量子力学，这就是创造了整个二十世纪的物理发展，就是越小越好，把这个物质大的小的，你把小的找着了，你就整个解决掉了。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    这个成功我等会把这个成功再重复一下，在二十世纪中叶就影响生物（科学）生物是宏观的，克利克，他是念物理的，他说我们要把生物也同样做，结果他就发现DNA，后来就是基因，也是觉得越小越好，就是现在的生物的大成功，也是跟了简化归纳，可是这个方向，在二十一世纪是另外一个方向，这是二十世纪的，也可以说是经典的一个方向，不过这个方向非常成功，它一个科学研究，我们总要知道一点，它跟浪头一样，一个浪一个浪地（涌）过去，当一个浪最高的地方，事实上是以后就下来了，所以你不能看人家这个浪在哪最高，然后你跑过去，它根本是人家另外一个浪过来了，所以这是一个一个，我们叫wave（波浪）出来的，这是二十世纪的，可是二十世纪很成功，它成功，刚才已经讲了，1900年1925年建立相对论跟量子力学，紧跟着就是二十世纪科学问题，原子构造，分子结构，核能、激光、半导体、超导体、超计算机，然后我们找到所有的核子、介子，它实际上所有我们知道的物质，是由12种（基本粒子组成），这是很厉害的，这个12种（基本粒子）叫夸克跟轻子两种，每个六样，夸克，六个夸克，上、下、奇、粲、顶、底，核子也是它这个做的，种子也是它做的，铀也是它做的，碳也是它做的，他说这六个是基本的，还有六个轻子，电子，u介子，t介子，电子中微子，u中微子跟t中微子，这个12个就组成我们所有我们了解的物质，这是二十世纪一个极大极大的贡献。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    质子、中子、核子不是重要的，它（重要的）是夸克，所有的作用，电作用，磁作用、弱作用，强作用，电磁作用一共就变成三大作用，强作用就是核子怎么做的，把夸克怎么联起来的，电磁作用，电弱作用，就是弱的β衰变，跟电，跟磁联在一块的，跟引力场，就是爱因斯坦（理论）就这三个大作用，我们把所有的我们了解的宇宙就是12种粒子，三个大的作用，这是很成功，可是不然，它有问题，试想想，二十世纪初十九世纪末，有两个大谜，我们要看将来不是要看我们的成功，是要问我们有什么是大的问题，那才会找到二十一世纪怎么发展，我认为二十一世纪有四个大问题，这四个大问题是什么呢？四个大问题是，我们这个三大作用的理论都是对称的，刚才讲了很多不对称，不过这个理论都是对称的，可是实验结果不对称，那你说这个问题是显然这个回答很简单了，把理论物理学都赶掉就完了，因为这显然不合理，不然，理论对称实验不对称有密切关系，这是现在的大谜，所有的基本粒子，12种，一般看不见，所有的夸克，单独都看不见，这也是一个大谜，90％以上的全宇宙物质不是我们的物质，是暗物质，我们是少数，有类星体，现在大概至少有100多万个，每个类星体的能量是太阳能量的10的15次方到16次方大，一个太阳能量核能跟煤的这个能就差了几万倍，十万倍，这是（10的）15次方，这远远远远超过了太阳能，不是我们知道的能力，这四个问题就是二十一世纪我们要进攻的问题。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    为什么理论对称，而实验不对称，为什么一般这个基本粒子我们单独不能看见，为什么绝大部分的物质是暗物质，为什么有类星体这种能量存在，那么这个里边，这两个，我只能告诉各位，暗物质的证明在哪，类星体的证明在哪，解还没有解出来，不过我们有一个可能解，然后我们再讨论，现在可能解的方面，我先把这四个大问题列出来，这四个大问题，你可能可以想这个可能还（可以理解），另外两个显然是很奇怪的了，就是说为什么对称跟不对称有什么关系？一般的基本粒子为什么看不见，暗物质从哪来的，类星体从哪来的，我们就先讲一下大体的我们的想法，理论对称实验不对称，它就跟宇宙开始大爆炸是对称，我们现在宇宙不对称，有密切关系，这个对称跟不对称，是一而二，二而一的事情，这是第一点。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    基本粒子一般看不见的道理是我们的真空在变，真空把夸克给禁闭起来，所以我们要改造真空，暗物质怎么来我们不知道，不过我们知道在宇宙大爆炸刚开始时，我们好像很了解，我要给你们一个了解，所以我们想制造近于宇宙开始大爆炸的情况，看到底这个是怎么出来的，也要看看到底类星体怎么出来的，所以我们基本的方法就是说，要制造跟宇宙大爆炸很（相）近的情况，在实验室，然后跟踪去，看它怎么发展，所以我现在先讲一下，就是对称跟不对称的关系，就是夸克为什么看不见，然后（讲）暗物质的证明是什么，然后（讲）类星体的证明是什么，然后（讲）我们有什么方法，当然（还有）这个二十一世纪的发展，我觉得是可能跟二十世纪是同样有很大的发展，怎么发展，当然我并不知道，所以我觉得做这个研究是很有意思的。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">李政道：</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">所以现在就开展讲这个对称跟不对称，为什么要相信对称，我们生活的世界，是充满着不对称，那我现在第一步就讲这个道理，我觉得是最高的对称性就是最多的不对称可能性，我现在举个例子，所以请各位稍微有点耐心，这个叫弯曲，怎么样弯曲呢，我身上带了一根杆子，这个杆子假如你加力在两边，它就缩下去，可是我的力加得忽然很快，它就弯了，这一点是不是大家很清楚，一根杆子，假如杆子力加得很快，它就要弯的，这个问题早在200多年前就解决，是欧拉解的，他说这个F等于力，假如F大于一个值的时候，这个弹性常数，转动惯性，杆的长度有关系，（根据）这个公式，它就弯了，我们并不要去证明欧拉这个定理，不过我要给各位看一下，你把这个杆子，这个杆子你拿来，假如你看它的截面，它截面可能是圆的，也可能是长方形的，也可能是最后这样弯的，圆的当然非常对称，各个方向都一样，是不是，长方形呢，它只有两个方向，一个沿着长方形，向一边，向另外一边，（弯）的是什么对称性都没有，假如圆的话，你看，我可以压坏，它可以这样弯，也可以这样弯，也可以这样弯，是不是很清楚，所以圆的呢，它是最对称，可是，不对称弯的时候，可能性是无穷大的，这是不是很清楚，假如这个截面是圆的话，力加大，它可以这样弯，也可以这样弯，这样弯，所以最对称方向，它有最多的无穷多的不对称的弯曲，假如它是长方形，就是一个不太好的长方形，那么我一压它可以这样变，可以这样变，可是这样不会的，这个对称性小了，随着小了，不对称的可能性也小了，就是说，本来原始常态的对称性越大，将来变成不对称的可能性也就越多，这是把对称跟不对称演变（得）出来的，所以这是我们要用这个例子就是说我们当初宇宙开始大爆炸时是绝对对称的，因为它绝对对称，所以我们可以说，（以后）很不对称，就联起来了，所以这是第一点。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    所以第二呢，我就把它总结一下，圆截面是具有最高的对称性，这是圆截面，因为它具有最高的对称性，所以它不对称的可能性是有很多很多，所以最高的对称性跟最多的不对称的可能性是相联的，这个（长方形）的没有这么多大的对称性，所以它的不对称也只有两个方向，是把对称跟不对称在它的运用的这边把它表达出来，第二个，假如说我们开始是绝对对称，我们现在不对称，我们怎么知道本来的情况是对称呢？这电就比较困难一点，就是本来是对称，是圆截面，现在不对称，已经（变）在这了，你怎么知道你切一刀，不准切，你怎么能推出来，它是这样，假如说它原来是圆的，它可以这样弯，像弯度一样的话，它的能量是一样的，因为它是对称的，所以可以从这个位置搬到这个位置，再动一点，不需要有什么激发能力它会动的，这个观点比较困难一点，这个就是现在我们要激发真空做这件事情。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    所以我再重讲一面，假如（它）本来是圆截面，有无穷多的可以弯，这些无穷多的弯实际上有关系的，因为它圆截面，你可以稍微动一点，就可以从一个弯的角度可以弯到这去了，从这到这，它的能量是不需要的，所以你就是在这不对称的位置你可以激发，用激发的方法你可以试验出它当初是怎么样，这个观念是一个很重要的观念，通过不同不对称弯曲的方向，使这个不同的是有转动的联起来，不需要能量的激发，这个我们叫Goldstone－Nambu玻色子，这个玻色子我们实验可以做，在粒子物理里面，这个弯曲的杆就是物质的真空，在宇宙开始大爆炸的时候是对称的，我们认为因为物理的真空在变，所以现在的真空态对CP也好，对C也好是不对称的，这个不对称的真空也可以夸克禁锢起来。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    所以我们第一步呢，是要把真空要改革，我们是有电子跟核子构成的，现在的宇宙是CP不对称的，因为这个不对称，我们才存在，假如对称我们就不会存在，为什么呢？假如对称的话，我们是负电子，质子做的，假如另外有一个是反物质，那这个物质跟反物质消灭掉了，变成能量了我们就不可能存在了，我们所以存在，就因为它没有反物质在我们旁边，假如有反物质存在，我们实验室很少数的，有控制的，所以我们认为从宇宙开始，它实际上就是一个能量在放大，可是它这个真空态从宇宙开始到我们现在不一样，那么真空怎么能变，什么是真空？真空是没有物质的态，可是力可以走得过，怎么呢？地球跟太阳之间很接近真空，地心引力显然走过去了，要不地球不会绕着太阳转，星云跟星云之间这是更接近真空，不过星云之间现在我们也可以算出它的吸力，所以力场是走得过的，力场走得过，万有引力走得过，强作用走得过，电弱作用走得过，有作用可以走得过，它就有涨落，所以真否有涨落，它不是以太它是和相对论性的，真空会因为它有很复杂的构造，有力走得过，虽然没有物质可以相变，可以禁闭夸克，可以把CP不守恒。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    然后呢，我解释怎么去（改造真空），我们现在研究很大的目标就要改造真空，在这个之前我再介绍一下除了一个是对称不对称是一个谜，夸克禁闭这是另外一个谜，我再要解释，暗物质跟类星体，暗物质的证明在哪呢，随便一个星云拿来，一个星云，你看得见，用眼睛看，光也可以，量它的辐射也可以，Y光，X光，不管（哪种）它都是一个星云都有个范围，普通相当于一千光年左右，你在这个星云外面，它总还有些，偶尔还有些（比如）一颗星了，但是（在）之间，这个星云它有地心引力在吸它，底地心吸力在吸它，假如（像）我现在地心吸力这样把它吸过来，它有东西呢，它就要转，（像）小时候，拿根绳子这样转，它转的速度跟吸力是有关的，实际上是转的速度的平方就跟吸力成正比，就是离心力了，所以量一个离开星云在这，这个幸运外面的一个灰尘也好，一个星也好，它的转动速度你可以量出，它受多少地心吸力，这一下在最近两次已经是几千个星云是我们银河系也可以，发现银河系在这，你到外面去，离开的距离差不多要十倍，它速度转的速度不下来，这就是你看，这是到这为止。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    请看下面一个星云，在我的有箭头的地方，这个上面坐标是速度，这是这个速度，这个速度，它不下来，看见没有，这是一个非常重要的一个事情，这是从这个红的底下的，就是这个红的底下的英文叫Luminous  mass，就是看得见的质量，在看得见质量这个之外，它那个转的速度不下来，它应该是没有地心引力了，地心吸力应该是越来越弱，越来越弱，它的速度就会下降，它不下来，那么这个不下来，很稀奇，这个叫NGC3192号，谁知道这是什么星云呢？银河系也一样，请看底下一个，底下一个。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    再后面一个，这是另外一个星云，这上面也是速度，这也是距离，它差不多是到的是在20的地方，是这个星云的位置，这个速度向上的，再请看另外一个，它这个星云这儿，可是在星云外面，这是速度，它的速度反而变高了，跟刚才的NGC（3192）一样的，再请看下面，每一个都这样，这是速度这是距离，差不多20，差不多一个星云的大小，它都不下来，没有一个例外，所有我们知道的星云，每一个都是这样，这就是什么意思，就是说，每一个（速度）都不下来，不下来甚至你看得见你用眼睛看也好，用X光也好，它还有东西在吸它，因此说，90％以上绝对的物质我们只能感受到地心吸力，不是我们的电子做的，也不是我们的反电子做的，也不是我们的核子做的，我们不知道是什么，这是宇宙的绝大多数，就在我们银河系，所以这当然是很惊人的，我们就是说，我们是宇宙的少数，我这里叫90％以上的暗物质，有人就说99％，就是，我们是绝对的少数，（这）绝无问题的，而是到处都（是暗物质），我们是少数，我们当然要了解它，咱们这个科学不能了解，你自己的音和星系，里面大部分的物质在哪里，所以我们这个方法，我们现在不了解，所以有一个方法，就是制造跟大爆炸很近似的时候，（看物质）到底怎么出来的，这是一个当代的一个大问题。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    另外是类星体，1961年最早是发现两个类形体，这是其中之一，剑桥273，它的能量是相当于太阳能（能量）这是很大的，10的16次方的能量，我们可以定可以看这个相片，可以定出它大小就跟地球跟太阳这个距离之间，所以它很小的，不是个星云，它比整个一个星云，整个星云里面，10的12次方的星，它（的能量）比那个差不多还大了一万倍，这是什么星，现在有差不多100万个，1961年最早的那时发现两枚，这是其中之一枚，在1982年2月就是这枚，一天能量增加一倍，这是绝对惊人的，你觉得好像不大奖励，忽然就增加一倍，太阳能的10的16次方，这是什么呢？</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    再看另外一个，这是我们叫爱因斯坦十字架，它是什么呢？这个（星）我们照像，最完美的一个，这个类星体是这四个（类星体）实际上是一个，是离我们的距离有80亿光年远，在它的近似10倍的距离，是这个星云，是另外一个星云，所以这是个星云，这是个类星体，这个类星体比中间这个星云是距离又多了十倍，它光走过中间那个星云时，它弯，所以中间的就等于现在像照相机一样，那么这个弯是这样走这样子，刚好这张相片有四个像，这四个像实际上是一个我们所以叫它爱因斯坦（十字架）爱因斯坦第一个提出相对论的，说光可以跟地心引力可以弯的，那么你这个地方你就可以看出来，这是中间吸的劲头，（从）这个劲头，你看这个类星体，距离差十倍，你看这光一点都不差，这是整个一个星云，这个类星体我们求出它的大小，是在地球到太阳距离之内，你用不着其他证据，你就知道它的能量了，它这样小的一个类星体它的能量可以，比这整个星云（还大）。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">李政道：</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    所以我们的宇宙是充满了不了解的挑战性的现象，不光是星云跟粒子物理，所以二十一世纪的物理发展是整体统一，是要把微观的基本里跟宏观的真空态一块研究，整体统一呢，我觉得是二十一世纪的发展，这个里面物理跟生物跟其他科学要平行发展，它为什么呢？因为精密计算，精密测算我们这套手法是跟生物的发展不一定完全一样的，他们当初DNA的发展就是跟物理的（研究方法一样），简化归纳出来的，整体统一也是要平行发展，那么各位就说，科学跟艺术有什么关系，我觉得这是很有关系的，艺术不管是诗歌，绘画、音乐是我们用新的方法唤起每个人的意识跟潜意识的情感。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    我们念李白的诗：“床前明月光”，就是你念的心里的时候，心里实际上是有的，你念了以后（觉得）它讲的真好，正是你心里想的情感，所以就说，情感越珍贵，唤起越强烈，反响越普遍，艺术越优秀，它是跨社会背景的，跨中国、外国、过去将来，有人类就有情感，它越普遍，越唤起，唤起的创造性艺术（越有）价值，人类的情感是每个人都有的，科学呢？科学不管天文、物理、生物、化学，对自然界的现象进行新的准确的抽象，科学家抽象的叙述越简单，应用越广泛，科学创造也就越深刻。</span><wbr /> <br><span style="font-size:16px;line-height:1.8em;">    自然界现在是，我们不在，自然界也在，可是这个抽象把它变成原理是我们做的，而我们把这个原理了解了，操纵了，我们也操纵宇宙了，所以呢，这个科学跟艺术它的共同基础呢，是人类的创造力，它追求的目的是真理的普遍性，它像一枚硬币的两面，是不可分割的，这是大家都很熟的《逍遥游》，这是庄子对宇宙的看法：“北冥有鱼，其名为鲲，化而为鸟，其名为鹏，鹏之大，不知其几千里也，其翼若垂天之云，背负青天，莫之夭阏”，庄子也挺有意思，他开头（说的）是鱼跟昆虫，后来是鸟，这跟我们进化论很象的，不过鹏之大，不知几千里也，鹏就等于整个宇宙，几千里对庄子（来）讲就是无穷大了，背负青天，莫之夭阏，没人可以抵达住，那么为了这个，刘巨德先生画了大鹏，这就是整个的宇宙，这就象征了宇宙开始的大爆炸，鹏的非就是宇宙开始的大爆炸，也象征了炎黄文化的过去跟将来。炎黄文化是有悠久的历史，就像宇宙开始大爆炸一样，会永远扩大，发展，勇往直前，谢谢大家！</span><wbr /> <!--v:3.2--> ]]></description>
<category><![CDATA[自然哲学]]></category>
<author><![CDATA[550115938@qq.com(wildyouth)]]></author>
<comments>http://550115938.qzone.qq.com/blog/1210483357#comment</comments>
<qz:effect>576</qz:effect>
<pubDate>Sun, 11 May 2008 05:22:37 GMT</pubDate>
<guid>http://550115938.qzone.qq.com/blog/1210483357</guid>
</item>

</channel>
</rss>

