ε Constrained Multi-mutant Rank-Based Differential Evolution Algorithm and Its Application in Multipath Repression

Many constrained optimization problems are encountered in engineering field. The problem of multipath repression in high-precision positioning navigation systems can also be regarded as a constrained optimization problem. Multipath repression can be transformed into a parameter estimation problem with a nonlinear state space model. In this paper, the multipath estimation problem in non-Gaussian noises is transferred into a constrained optimization problem. Then, a modified epsilon-constrained multi-mutant rank-based differential evolution (epsilon CMRDE) algorithm is explored to solve the optimization problem. For the proposed algorithm, the generalized maximum correntropy criterion (GMCC) is used as the objective function of the formulated optimization problem. The second-order moment of the estimation error and the prior information are taken as constraints to reduce the mean of the estimation error. Simulation results show that the proposed algorithm outperforms epsilon RDE algorithm in terms of global optimization ability and multipath estimation accuracy.