IEEE Access (Jan 2024)
Research on Thrust Allocation for Power Positioning System Based on Improved Sparrow Search Algorithm With Multi-Strategy Fusion
Abstract
This paper proposes an improved sparrow search algorithm (ADGSSA) based on multi-strategy fusion, and studies and verifies it in the thrust distribution of the dynamic positioning system of the new research vessel of Dalian Maritime University. First, the sparrow search algorithm is introduced. Aiming at its shortcomings in convergence speed, optimization accuracy and easy to fall into local optimality, a multi-strategy fusion improvement scheme including piecewise nonlinear dynamic random convergence factor strategy, warning position update condition and formula adjustment strategy, and golden sine perturbation strategy is proposed. Subsequently, the improved algorithm is applied to 12 benchmark functions with different characteristics and compared with Coati optimization algorithm (COA), whale optimization algorithm (WOA) and sparrow search algorithm (SSA). Experimental results show that ADGSSA outperforms the other three algorithms in optimization accuracy, convergence performance and stability. In the evaluation of 12 benchmark functions, ADGSSA demonstrates superior performance compared to SSA, WOA, and COA. Specifically, ADGSSA achieves the closest optimal value in 83% of the functions, provides the best average value in 85% of the functions, and has the lowest standard deviation in 75% of the functions. In the actual application of the thrust distribution system, ADGSSA shows significant advantages. Compared with WOA, the average power of the thruster is reduced by about 75%; compared with SSA, it is reduced by about 35%; compared with COA, it is reduced by about 17%. In addition, in 12 sets of thrust tests, the thruster wear of ADGSSA reached the smallest proportion of 82% among the four algorithms, which is significantly better than other algorithms. The test results and convergence curves further show that ADGSSA has significant effectiveness and superiority in improving the thrust distribution performance of the dynamic positioning system, proving the potential for promotion and application of multi-strategy fusion methods in control strategies.
Keywords
