A Self-Adaptive Multi-Learning Strategy Particle Swarm Optimizer for UWB Indoor Positioning Anchors Layout Optimization

Particle swarm optimization (PSO) is a widely used method for solving single-objective optimization problems. However, PSO suffers from diversity loss and premature convergence, leading to suboptimal performance in complex problems. To address these limitations, this paper introduces a novel self-adaptive multi-learning strategy PSO (SMLS-PSO) designed for real-parameter optimization tasks. SMLS-PSO integrates four distinct learning strategy-based PSO variants into a behavior pool and employs a new self-adaptive strategy selection mechanism. This mechanism dynamically chooses the most suitable learning strategy to update the velocities and positions of particles based on fitness information and payoff at different evolutionary stages. To enhance the performance of SMLS-PSO, three key improvements are incorporated: (1) a stagnation counter parameter to minimize wasted fitness evaluations; (2) a boundary symmetry mapping method to manage out-of-range searches; and (3) a Quasi-Newton local search operator to boost local exploitation capabilities. SMLS-PSO is first compared with four component PSO variants on 13 basic benchmark functions. Subsequently, SMLS-PSO is compared with eight state-of-the-art PSO variants on both 30D and 50D CEC2017 test suite problems. Experimental results indicate that SMLS-PSO statistically and significantly outperforms the compared algorithms on the majority of the test problems, showcasing its superior optimization capabilities. Finally, SMLS-PSO was applied to the optimization problem of UWB anchors layout, and it achieved a higher locatable space coverage rate and a better average HDOP value compared to the conventional layout scheme and other PSO variant optimization schemes.