改进的多任务分配与调度遗传算法

Improved Genetic Algorithm of Multi-task Matching and Scheduling

提出了一种改进的混沌遗传算法,定义了问题模型,采用一任务序列加一主机序列的编码表示,并由约束关系阶段随机产生合法调度。为确保生成新调度,改进了交叉机制;为保持解的多样性和均衡负载,设计了任务均衡变异算子。为利用混沌优化技术,交叉、变异频率都由混沌序列动态控制。仿真结果表明,该算法总能得到最佳解。

An improved chaos genetic algorithm theory was proposed. The problem model was defined. The chromosome encoding was represented by a task sequence and a host sequence. The valid scheduling was assembled by randomly selecting host and task of each stage task set in turn. To ensure generate new valid scheduling, the crossover mechanism was improved. To balance load and increase solutions diversity, tasks balance mutation operator was designed. To use chaos advantage, crossover and mutation frequency were dynamically controlled by chaos iteration sequence. The simulation results showed that this algorithm always get the best solution.