This chapter builds on the notions presented in the previous Chap. 2 about search spaces. Here we adopt a problem instance-based point of view in the sense that instance search spaces are examined using a number of statistical methods and well-defined metrics. This empirical analysis is based on sampling the space. The results give information about the hardness of a given problem instance and potentially enable clever techniques that exploit the search space structure in an algorithm in order to find a quality solution to the corresponding optimization problem in reasonable time.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Statistical Features and Metrics of Search Spaces

  • Bastien Chopard,
  • Marco Tomassini

摘要

This chapter builds on the notions presented in the previous Chap. 2 about search spaces. Here we adopt a problem instance-based point of view in the sense that instance search spaces are examined using a number of statistical methods and well-defined metrics. This empirical analysis is based on sampling the space. The results give information about the hardness of a given problem instance and potentially enable clever techniques that exploit the search space structure in an algorithm in order to find a quality solution to the corresponding optimization problem in reasonable time.