摘要： 提出一种综合考虑Agent时间片和执行截止期限要求的CPU时间片组合拍卖遗传算法。该算法定义了问题模型，采用可以去除不具备竞争力标的预选择策略，减少遗传算法的计算复杂度。在遗传算法求解过程中，设计适合该问题的变长染色体编码方式、交叉算子、换序算子和换标算子。仿真实验结果表明，该算法取得了符合要求的最优解。

关键词: 组合拍卖, 遗传算法, 截止期限, CPU时间片

Abstract: This paper proposes a Genetic Algorithm(GA) for CPU time slice combinational auction in which both Agent time slice demand and execution deadline demand are considered. The problem model is defined, and the preselecting method is adopted to filter non-competitive bid to decreasing computing complexity of genetic algorithm. Variable chromosome coding, crossover operator, adjusting-bid-order operator and replacing-bid operator are designed to solve the problem in genetic algorithm. Experimental results show that this algorithm can obtain satisfaction optimal solution.

Key words: combinational auction, Genetic Algorithm(GA), deadline, CPU time slice

中图分类号: 

• TP391