论文部分内容阅读
并行测试的任务优化调度是并行测试技术的核心问题。为了解决现有调度方法耗时、实际应用范围有限以及缺少对资源冲突和系统死锁的形式化分析等问题,采用赋时有色Petri网(TCPN)建立并行测试任务调度的TCPN模型,基于TCPN模型的可达标识图利用改进蚁群算法求解最优任务调度序列。算法搜索过程中,采用多目标优化,目标函数综合了测试时间、仪器成本和负载平衡度,使得算法更符合工程应用。采用动态标注方法在搜索过程中加大可行解间的信息素差别,避免算法早熟。仿真实例证明该算法是有效的。
Task scheduling in parallel testing is a core issue of parallel testing techniques. In order to solve the problems of time-consuming, limited practical application, and lack of formal analysis of resource conflicts and system deadlocks, the TCPN model of concurrent test task scheduling is established using timed colored Petri nets (TCPNs). Based on the TCPN model The reachable identification map uses the improved ant colony algorithm to solve the optimal task scheduling sequence. In the algorithm search process, the multi-objective optimization is adopted, and the objective function combines the test time, instrument cost and load balancing degree to make the algorithm more suitable for engineering application. The dynamic annotation method is used to increase the pheromone difference between feasible solutions in the search process and avoid the algorithm premature. The simulation example proves that the algorithm is effective.