Combinatorial optimization and simulation for weapon system portfolio using self-adaptive Memetic algorithm

The weapon system portfolio problem is considered as a typical constrained combinatorial optimization problem with the purpose of maximizing the expected damage of hostile targets. considering the computation complexity and the strict time constraints,a decision-making methodology based on self-adaptive memetic algorithm is proposed as an alternative to help military commanders in making appropriate decisions. in this framework,self-adaptive genetic algorithm performs global search to prevent trapping into the local optima,where the crossover probability and mutation probability could be adjusted dynamically according to the prematurity degree of evolving population. furthermore,problem-specific heuristics are utilized to conduct local search and fine-tuning in the solution space. a case study is given to illustrate the entire procedure and verify the performance of our proposed algorithm. comparative experiments show that our algorithm outperforms its competitors with regard to solution quality and computation time. in addition,very large-scale scenarios are also simulated to demonstrate the scalability of our algorithm. © 2017,university of kuwait. all rights reserved.