针对分布式综合模块化航空电子网络分区方法,指出其对消息传输模式配置存在不确定性;通过形式化描述将问题转化为包括实时性约束、带宽约束、缓存约束的最优化问题;提出了基于遗传模拟退火的传输模式配置算法,以系统的消息端到端延迟均衡为优化目标-通过给出两个具体算例对算法有效性进行了验证,同时比较了本算法与传统遗传算法的性能-对比结果表明,本算法能够解决传输模式的配置问题,尽管计算时间较传统遗传算法高出18-1%,但所得到的适应度值高出28-7%-本算法为网络分区在实际航电系统中的应用提供了参考-
It was indicated that configuration of message transmission mode among network partition meth-od in distributed integrated modular avionics was uncertain- Through formal description the problem was trans-ferred into an optimal problem with constrains including real-time, bandwidth and buffer- Then the transmis-sion mode configuration algorithm was proposed based on genetic simulated annealing algorithm; the objectivefunction was aimed to balance of end-to-end delay of system messages- Effectiveness of the algorithm was veri-fied based on two experiments, and performance of the algorithm was compared with that of traditional geneticalgorithm- The results show ahhough the algorithm is 18-1% slower, the fitness value is 28-7% higher- Thisalgorithm provides a reference for the network partition application in avionics systems-