• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Knowledge Technology & Policy >Parallel Experimentation and the Problem of Variation
【24h】

Parallel Experimentation and the Problem of Variation

机译：并行实验和变异问题

获取原文
获取原文并翻译 | 示例
       
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Evolutionary processes have two underlying mechanisms: variation (expanding options) and selection (narrowing options). Darwin and Darwinism focused on the selectional mechanism, i.e., natural selection. If an evolutionary process is seen as pushing towards higher altitudes on a multi-peaked and cloudy "fitness landscape," then selection is a hill-climbing mechanism while a mechanism for variation would explore to find higher hills. Selection without variation might get stuck on a low hill. Our thesis is that the best solution for the problem of variation is parallel experimentation with the different probes having some semi-isolation from selectional pressure and where results in the experiments can be cross-communicated and compared to ratchet up the performance of the whole group. Our approach is by seeking out the common elements addressing the problem of variation in evolutionary theory (Sewall Wright's shifting balance theory), in various search procedures, in the selection of economic projects, and in science and technology development. The opposite strategy is series experimentation which assumes that one is already climbing the highest peak so consideration of other peaks is irrelevant. The main critical applications are to global institutions such as the World Bank or IMF who do not condone parallel experimentation since they already "know" what is the highest peak—and, more generally, to those in power on top of a low hill who know that any real variation would be downhill for them.
机译：进化过程具有两个基本机制：变化（扩展选项）和选择（缩小选项）。达尔文和达尔文主义侧重于选择机制，即自然选择。如果在多峰且多云的“健身景观”上进化过程被视为朝着更高的海拔方向前进，那么选择是一种爬山机制，而通过变化机制来寻找更高的山丘。没有变化的选择可能会陷入低谷。我们的观点是，解决变异问题的最佳解决方案是使用不同压力的探针进行平行实验，这些探针与选择压力具有一定的半隔离性，并且可以交叉交流实验结果并进行比较以提高整个研究组的性能。我们的方法是寻找解决进化论（Sewall Wright的转移平衡论），各种搜索程序，经济项目的选择以及科学和技术发展中的变异问题的共同要素。相反的策略是串联实验，它假设一个已经在爬上最高峰，因此与其他峰的考虑无关。主要的关键应用是对诸如世界银行或IMF之类的全球机构，他们不容忍并行实验，因为它们已经“知道”什么是最高峰，而且更普遍地，对于那些在低矮山丘上掌权的人，他们知道任何真正的变化对他们来说都是下坡路。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号