移动Agent多任务调度算法

提出一种可覆盖全部解空间的移动agent多任务分配与调度混合遗传算法。给出问题模型及染色体表示方法，采用禁忌表加随机算法生成初始种群，设计新的交叉机制保证交叉进化解的合法性。为促进算法的收敛，变异个体使用禁忌及任务均衡启发变异算子。还采用保持解的不降性的最佳个体保留策略。2种任务节点、3种通信代价、3种主机节点共18组图的仿真结果表明该算法进化的最优解较标准遗传算法有37.1%的平均改进量。

Mobile Agent Multi-task Scheduling Algorithm

A hybrid genetic algorithm is proposed to search the optimal solution of mobile agent multi-task matching and scheduling problem. The problem model and chromosome representation are defined. The initial population is generated by the taboo search and random selecting method. And a new crossover mechanism is designed to make the new scheduling evolved by crossover mechanism valid. To accelerate the algorithm convergence, taboo search and tasks load mutation operator is adopted. Also the best chromosome preserving strategy is adopted to keep optimal solution non-decreasing performance. Simulation results of 18 graphs combined by 2 kinds of task nodes, 3 kinds of communication costs and 3 kinds of host nodes show that the algorithm can get 37.1% improvement compared with standard genetic algorithm.