作者: Enver Yücesan , Sheldon H. Jacobson
关键词: Hill climbing 、 Negative probability 、 Search problem 、 Heuristics 、 Algorithm 、 Convergence (routing) 、 A priori and a posteriori 、 Mathematics 、 Discrete event simulation 、 Machine learning 、 Mathematical optimization 、 Event (probability theory) 、 Artificial intelligence
摘要: The search problem, ACCESSIBILITY, asks whether a finite sequence of events can be found such that, starting with specific initial event, particular state reached. This problem is intractable, indicating the need for heuristics to address it. One difficulty when applying ACCESSIBILITY assessing priori their effectiveness, and knowing how best adjust them improve performance. paper introduces false negative probability as performance measure generalized hill climbing algorithms applied discrete optimization problems, using analysis framework. also used obtain necessary convergence conditions. implications these results on GHC effectively are discussed.