中文摘要：

为满足复杂装备中某些关键系统的实时测试与诊断需求，在测试性设计过程中开展并行测试调度优化研究；针对测试任务之间关联关系复杂的实际情况．通过建立基于图染色理论的测试任务关系模型，实现了对系统资源冲突、死锁等问题的形式化描述，将并行测试调度优化转化为求解图的色数问题；并在求解图的色数问题过程中，利用改进的遗传算法逐步求解图的最大独立集，即可并行测试的测试任务子集，从而获得并行测试调度方案；最后，利用某实例对文章提出的并行测试任务调度建模及优化方法进行了验证，优化效果明显，证实了该方法对解决并行测试调度优化问题的有效性。

英文摘要：

In order to satisfying real--time test and diagnosis in some key system, it needs to make research on optimization of parallel test scheduling in the process of design for testability. For the complicated relationship among tests, it first built a relation model based on Graph Coloring Theory （GCT） which can describe the dependency relationship among tests. By this way, the problem of system resource conflict and deadlock can be formally described, and then the optimization of parallel test scheduling is transformed to be the problem of graph chromatic number. In the process of getting the chromatic number of the graph, we use improved Genetic Algorithm （GA） to get max inde pendent set step by step, which is test subset that can he executed in parallel, and then getting parallel test scheduling scheme. Finally, it verified the method of modeling and optimizing for parallel test scheduling by using an example, the optimization effect is good. It proves that this method can be applied in solving parallel test scheduling effectively.