摘要
为有效求解任务分配问题,提出带有推荐功能的蚁群算法。构建了一种推荐机制,根据对问题的分类情况,基于蚁群算法的算子规则与问题的匹配程度,为每类具体问题的求解提供算子推荐。为提高算法的求解性能,针对问题的三个优化目标设计了三种局部搜索策略,在蚁群算法迭代过程中,根据解的迭代特性自适应地嵌入算法中执行。设计了四种类型共16个不同规模的算例来验证方法的有效性,通过验证每类算例在不同规模下算子规则选择的一致性,从侧面反映了算法推荐机制的合理性。
To effectively solve the task allocation problem,an ant colony algorithm with automate recommendation was put forward.According to problem classification,an recommendation mechanism was constructed to recommend operators for each specific problem based on operator rules and matching degree.To improve the solution performance,three local search mechanisms were designed aiming at the optimization objectives.The local search was adaptively embedded based on the quality of each index in the process of iteration.16 instances of different scales with four types were generated to prove the effectiveness of the recommendation mechanism.Consistence of operation rule selection under different scales of each tyoe was used to exolain the rationality of the mechanism.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2013年第9期2220-2228,共9页
Computer Integrated Manufacturing Systems
基金
国家自然科学基金资助项目(71031007
71101150
71071156
61203180
71101013)
国防科技大学优秀研究生创新资助项目(S120501)~~
关键词
任务分配问题
蚁群算法
算子推荐
局部搜索
task allocation problems
ant colony algorithm
operators recommendation
local search
