作　　者： ; ;

机构地区： 东北大学计算机科学与工程系

出　　处： 《自动化学报》 1994年第2期235-239,共5页

摘　　要： 本文将函数逼近的方法引入到学习式搜索中，使学习式搜索通过一定数量的解题训练后可以建立起一个任意一致逼近理想函数ｈ＊（·）的启发估计函数ｈ（·）．本文给出了一个这样的学习式搜索算法Ａ－Ｂｎ，并证明了当训练例子集充分大后，Ａ－Ｂｎ可在多项式复杂度内解决任一后来提交的同类问题． In this paper,Polynomial Approximation method and theory are introduced into the research of Learning Search of Artificial Intelligence.In this way, we can use a search algorithm repeatedly to construct a heuristic estimate function h(·) which uniformly approximates to the optimal estimate function h*(·) with arbitrarily high precision. One of such learning setrch algorithms, A-Bn is presented and it is shown that, when the number of training samples becomes large enough, the worst-case complexity of A-Bn can be reduced to O(poly(N)),where N is the length of the optimal solution path, poly (N) is a polynomial of N.

关 键 词： 人工智能 函数逼近 学习式搜索

领　　域： [自动化与计算机技术] [自动化与计算机技术]