多目标蚁群算法及其在固液混合火箭发动机系统优化设计中的应用

A Multi-Objective Ant Colony Algorithm and Its Application in System Optimization Design for Hybrid Rocket Motor

为了提高求解多目标优化问题的Pareto解集的效率,建立了适用于多维、多目标优化问题的多目标蚁群算法(Multi-objective Ant Colony Algorithm,简称MACA)。该算法首先修正了蚁群算法的信息素更新机制和转移概率,然后改进了蚂蚁的行进策略,即提出了依概率选择搜索策略。最后,应用该算法对某型号固液混合火箭发动机系统进行了优化设计。计算结果表明,多目标蚁群算法获得的Pareto解集分布均匀、散布范围广,可以有效解决多目标优化问题,能为决策者进行目标权衡提供充分依据。

In order to improve the searching efficiency and keep the diversity of Pareto set,a modified ant colony algorithm called as multi-objective ant colony algorithm（MACA） is built for the high-dimensional and multi-objective optimization problem.For the nature of the multi-objective optimization problem,two moving strategies are employed in the searching process to ensure better solutions,one is modifications of the pheromone update mechanism as well as the transition probability,and other is modification of the moving strategy in probability.As a specific example,the MACA is applied to the system optimization design of a certain hybrid rocket motor.The optimization results show that present method is effective and efficient for multi-objective optimization problem,and facilitates the decision maker to choose the best design with the tradeoff.