云制造环境下面向多目标优化的虚拟资源调度研究

为解决云制造环境下虚拟资源调度存在的算法求解效率不高、模型建立缺乏考虑任务间关系约束和任务间及子任务间的物流时间及成本因素等不足，构建了兼顾交货期时间最小化、服务成本最低化、服务质量最优化为目标的多目标虚拟资源调度模型；采用一种基于项目阶段的双链编码方式进行编码，并提出自适应交叉与变异概率公式，以避免交叉、变异概率始终不变导致算法效率下降与过早收敛的问题；在此基础上利用基于项目阶段的多种交叉变异策略相结合的改进遗传算法进行求解，保证了算法的全局与局部搜索性能。实例结果表明，相比于传统的模型与算法，该模型适用性更强，改进的遗传算法在求解效率、准确度与稳定性方面均有较大提高。

Research on virtual resources scheduling for multi-objective optimization in cloud manufacturing environment

In virtual resources scheduling of cloud manufacturing, algorithms were often inefficient; relational constraints between tasks, as well as logistics time and cost between tasks and sub-tasks, were not fully considered while modeling. In order to solve these problems, this paper constructed a multi-objective virtual resources scheduling model with the goal of minimizing the delivery time, minimizing the service cost and optimizing the quality of service. We used a double-stranded coding method based on project stage, and proposed an adaptive crossover and mutation probability formula, so as to avoid the low efficiency or premature convergence of algorithm caused by the invariance of crossover and mutation probability. On this basis, to guarantee the global and local search algorithm performance, we used an improved genetic algorithm combined with multiple crossover and mutation strategies based on project phase. The example shows that our model is more applicable than the traditional ones, and the improved genetic algorithm has considerable improvement in efficiency, accuracy and stability.