A Cooperative Neural Network For Constrained Multi-objective Optimization With Application To Digital Filters

Multi-objective optimization（MOP） has widely used in various fields such asfinance, engineering, and science. Different from traditional single objectiveoptimization, MOP aims at finding the optimal solution with best trade-offs betweenthese objective functions. This paper first presents a new cooperative neural networkfor solving constrained multi-objective optimization problems. The constrainedmulti-objective optimization problem is reformulated into two constrained singleobjective optimization problems and two neural networks are designed to obtain theoptimal weight and the optimal solution of the two optimization problems,respectively. The proposed algorithm has a low computational complexity and is easyto be implemented.Digital filter is a very important technology for digital signal processing, which hasbeen widely used in speech signal processing, image signal processing and biologicalsignal processing. The design of digital filters involves simultaneous approximationdesign in frequency and time domains, which is stated as a multi-objectiveoptimization. Finally, the proposed algorithm is well applied to the design of digitalfilters. Computed results illustrate the good performance of the proposed algorithm.