A modified ant colony optimization (ACO) metaheuristic is integrated with a local search procedure to produce an algorithm that combines the ability of ACO to locate global optimum solutions with the local search intensification ability of Nelder and Mead's simplex method to efficiently locate local optima within the solution space. Optimizer performance is evaluated according to two criteria : (1) "fitness function" accuracy, or how closely solutions are able to match a given "objective function", and (2) convergence speed, which is based on how many calls to this function are required in order to reach the optimal solution found by the algorithm. Design of a star grain solid rocket motor (SRM) to match a specified chamber pressure vs. time curve is used as the objective function. The SRM performance model, supporting background information, algorithm details, and test results are described and future work possibilities are discussed.
展开▼