Enhancing Battery Electric Vehicles Safety Optimization With Adaptive MOEA/D Based on Dynamic Grid and Multiple Dominance

With the rapid development of electric vehicles (EVs), the safety of battery systems has become an increasing concern. However, the application of multi-objective optimization algorithms in the safety domain of EVs is rare, particularly in optimizing complex, non-linear multi-objective problems (MOPs) such as crashworthiness and thermal management. This work proposes an improved multi-objective evolutionary algorithm based on decomposition (MOEA/D), which introduces a dynamic grid system for building stability matrices to determine whether the current population reaches a stable convergence state with less computational demands and designs a novel sparsity level evaluation to measure the weight vectors. Additionally, we propose a multiple dominance to address the Pareto dominance limitation. These aim to optimize the key issues in EV safety. To further demonstrate the performance of our proposed algorithm, we compare it with 10 state-of-the-art algorithms on 26 benchmark problems. The experimental results based on multiple performance metrics show that the proposed algorithm have outstanding performance. Note to Practitioners—Ensuring the safety of the battery system in EVs under collision conditions is important. This paper proposes an improved MOEA/D, aiming to optimize the maximum energy absorption of battery protection materials, enhance crash force efficiency, and improve the temperature difference of the battery system coolant. In proposed improved MOEA/D, the dynamic grid system and sparsity level evaluation ensure the acquisition of feasible solutions that align with users’ diverse preferences. Additionally, the multiple dominance strategy further enhances the quality of the solutions, leading to more superior optimization results. Experimental results demonstrate that the proposed algorithm shows outstanding optimization performance. We also further analyze the influence of each DAH structural parameter on performance, which enables users to make a trade-off according to their specific application requirements. Future research will explore the interdisciplinary integration of battery protection optimization to achieve more refined and precise optimization.