摘要
由于模拟生物免疫系统功能的免疫算法具有解决复杂工程问题的潜力,同时也存在容易陷入局部最优平衡态和进化后期搜索停滞不前的缺陷,在深入分析生物免疫系统机理之后,将生物免疫原理和生物遗传理论集成到免疫算法中,提出了双倍体免疫算法。这种算法采用疫苗、双倍体等多种生物机制,不仅防止了早熟,而且加快了收敛速度。最后通过典型的Job-Shop调度Benchmark问题LA21的求解证实了此算法的有效性和可行性。
Immune algorithm （IA） which simulates the biology immunity system has potential to provide novel method for solving the complicated engineering problem. However, immune algorithm inevitably has two serious limitations, that is, running easily into local optimization and search stagnancy in last phase. After biology immunity mechanism is thoroughly analyzed, diploid immune algorithm （DIA） is put forward based on the integration of biology immunity principle and biology genetic theory. The algorithm adopts bacterin, diploid and other biotogical mechanisms, which can not only avoid prematurity but also accelerate constringency. Finally, the effectiveness and feasibility of the algorithm are validated by simulation experimentation of LA21 which is a typical Job-Shop Benchmark problem.
出处
《工业工程》
2007年第2期93-96,101,共5页
Industrial Engineering Journal
基金
国家863/CIMS主题资助项目(2003AA411110)
航空科学基金资助项目(01H53061)
教育部博士点基金资助项目(20040699025)
