冗余系统费用模型及优化

Cost Models and Optimization of Redundant System

根据冗余设计的不同形式 ,在Tillman并联系统费用模型的基础上引入了系统功能指标 ,并将其推广至其它形式的冗余系统 ,从而得到了一个较为完整的冗余系统费用模型。在此基础上 ,建立了用于优化并串联冗余系统设计的非线性混合整数规划模型 ,分别使用了模式搜索与启发式算法的组合算法以及遗传算法对其进行求解 ,着重讨论了两种方法的实现关键及各自的优劣。最后 ,经计算比较 ,提出了遗传算法与模式搜索、启发式算法相结合的求解方法 ,并说明了该方法的有效性。

According to the different styles of redundancy, the index of system function is introduced into the Tillman's cost model of parallel system, which is then extended to other forms of redundant systems. As a result, an integrated cost model of redundant system is achieved. Based on this model, nonlinear mixed-integer programs are built to optimize the design of the parallel-serial system. Pattern search-heuristic algorithm and genetic algorithms are used respectively to solve the problem, and the key implementation points as well as advantages and disadvantages of the two methods are discussed with emphasis. In the end, through calculation and comparison, a method of combining genetic algorithm, pattern search-heuristic algorithm is put forward, and its effectiveness is illustrated.