基于蚁群算法的固体火箭发动机总体参数优化

System parameter optimization of solid propellant rocket motors based on ant colony algorithm

为建立一种支持连续域、离散域混合变量的优化算法以用于固体火箭发动机总体参数优化,改进了基本蚁群算法,融入"网格划分"、"哑元化"和"变尺度局部搜索"三种策略,以改进算法的寻优性能和使用范围,其中局部搜索算法仍采用蚁群算法。使用了几个较具欺骗性的经典测试函数对改进蚁群算法进行了测试,计算结果表明改进蚁群算法找到全局最优值的概率较大。应用改进蚁群算法对固体火箭发动机总体设计中的两个重要总体参数——燃烧室工作压强和喷管面积比,进行了优化求解,获得了满意结果。诸算例的优化结果表明,该改进蚁群算法具有支持混合变量,全局寻优性能稳定和搜索精度高的优点,对工程优化设计问题具有较好的寻优性能和更强的适用性。

In order to construct an optimization algorithm supporting continuous and discrete mixed-variables for system parameter optimization of solid propellant rocket motors, three strategies, mesh strategy, local search and dummy strategy were merged into basic ant colony algorithm to improve optimization performance and search precision, where the local search algo- rithm itself was still an ant colony algorithm. Several classical optimization algorithm testing functions, which are very entrapping were tested to verify the performance of the algorithm. The results indicate that probability is high for the algorithm to find global optimum value. The improved algorithm was applied to solve the two important system parameters, combustion chamber pressure and nozzle expansion ratio, of solid propellant rocket motor system design, and the results are satisfactory. All the results indicate improvements in three aspects, which are supporting mixed-variables, steady global optimization performance and higher search precision, respectively. Thus, the improved ant colony algorithm presents better optimization performance and better adaptability to engineering optimization design problems.