关键词: 半导体最终测试, 人工蜂群算法, 启发式规则, 贝叶斯网络, 多策略融合, 概率模型, 排序模型

Abstract:

This paper proposed an enhanced artificial bee colony algorithm (EABC) to minimize the maximum completion time of the semiconductor final test scheduling problem (SFTSP). EABC used hybrid heuristic methods to initialize the population, and used forward interpolation decoding strategies to improve the quality of the initial solution. A variety of exploration strategies based on the problem features and a Bayesian network?based probability model were designed to conduct a depth and width search of the problem solution space. In addition, a population update mechanism based on the restart strategy was proposed to strengthen the algorithm's ability to jump out of the local optimum. In the experimental part, multiple comparison algorithms were constructed to verify the effectiveness of the EABC’s key components. Simulation results based on some instances and comparisons with NFOA(Novel Fruit fly Optimization Algorithm)、KMEA(Knowledge?based Multi?agent Evolutionary Algorithm) and CCIWO(Cooperative Co?evolutionary Invasive Weed Optimization) demonstrated the effectiveness and robustness of the proposed algorithm.

Key words: semiconductor final test, artificial bee colony algorithm, heuristic rules, Bayesian network, multi?strategy integration, probability model, permutation?based model