Design of an Optimized Architecture for Manned and Unmanned Combat System-of-Systems: Formulation and Coevolutionary Optimization

With the rapid advancement of unmanned combat systems and the maturation of unmanned swarm technology, integrating unmanned combat systems into existing combat system of systems has become a hot topic. However, the current research at the organizational level is basically nonexistent. Therefore, we propose a methodology for designing an optimized architecture that is a significant product at this level to fill this technical gap. The design process consists of two steps: architecture modeling and coevolutionary optimization. The architecture model is composed of three basic elements, which are mission, equipment, and the command and control structure. The selection of the best scheme from possible solutions is a multiobjective optimization problem, and its computational complexity increases rapidly with the increase in the scale of combat system of systems. To overcome this difficulty, we introduce a decomposition-based evolutionary algorithm NSGA-III and improve it by using preference vectors to enhance the algorithm's local search capability. Finally, we conduct the comparative experiments on benchmark test suites and design an operation case to demonstrate the advance and effectiveness of the enhanced algorithm. The proposed architecture modeling and optimization method can achieve a set of nondominated solutions and provide auxiliary decision-making information to help commanders make decisions and arrangements.