摘要
为有效解决复合并行机排序的极小化最大完成时间问题,提出了分支定界算法和改进的启发式动态规划算法。利用分支定界算法的3个工具:分支模型、边界和优先规则,构建出分支搜索树。按优先规则进行定界搜索,从而减小了问题求解规模。将原始作业转换为虚拟作业,根据Johnson法则,求解出原问题的最优排序。改进的动态规划算法复杂度分析和计算实验表明,这两个算法可靠性高并且可以解决实际问题。
To minimize the Maximum Completion Time of one job scheduling model on complex parallel machines,a BB algorithm and a heuristic dynamic programming algorithm are proposed.A BB searching tree is constructed with three tools： BB model,boundary and priority rules.Priority rules are used to bound search,so as to limit the potential solutions’ boundary.Initial jobs are converted to virtual ones which meet Johnson Rule’s application conditions.By this way,using Johnson Rules,a solution of initial problem is achieved.These two algorithms are compared by complexity analysis and computational experiments.The result shows that both algorithms are reliable,and can be used in practical problems.
出处
《计算机工程与设计》
CSCD
北大核心
2011年第4期1311-1314,共4页
Computer Engineering and Design
关键词
复合并行排序
分支定界算法
生产调度
动态规划
启发式算法
scheduling multi-processor task
B＆B algorithm
production scheduling
dynamic programming
heuristic algorithm
作者简介
于蓉（1984-），女，山东济南人，硕士研究生，研究方向为嵌入式系统；
赵梗明（1959-），女，福建福州人，硕士，副教授，研究方向为嵌入式系统。E-mail：yr．xc@163．com
