摘要
集装箱装载是一个空间优化分解的布局问题,其约束条件多,属于典型的NP完全问题,求解难度大。在考虑实际应用中的约束条件下,使用三空间分割的布局方法,并结合分布估计算法(EDA)求解多约束装箱问题。同时对所使用的单变量边缘分布算法(UMDA)进行改进,采用了精英种群的策略并且加入遗传算法中的变异操作,这样能够使算法跳出局部最优解,加快算法收敛速度。实验结果表明该算法在求解速度和成功率方面都有明显的改善。
Container loading is a layout problem with space optimization and space decomposition. With multi ple constraints, it＇s a typical NPcomplete problem and difficult to obtain an optimal solution. Considering some constraints in practical applications, the measures of threespace dividing combined with the estimation of distribu tion algorithm （EDA） has been adopted to solve the multiconstrained pacing problem. At the same time, the uni variate marginal distribution algorithm （UMDA） by using elite population strategy and mutation of genetic algorithm is improved, which can get out of local optimal and accelerate the convergence of algorithm. The experimental re sults prove that this algorithm makes significant improvements on the speed and success rate of solution.
出处
《科学技术与工程》
北大核心
2014年第11期216-220,共5页
Science Technology and Engineering
基金
人工智能四川省重点实验室开放基金(2012RYJ04)
"青蓝工程"(苏教师[2010]27号)资助
