Decomposition-based interval multi-objective evolutionary algorithm for patrol robot path planning in chemical plants

Patrol robot path planning in chemical plants requires simultaneous consideration of path efficiency and operational safety under uncertain hazard conditions. Existing methods often model environmental risk in a deterministic manner, which may be insufficient for describing the uncertainty of hazard related risk in chemical plant environments. To overcome this limitation, this paper formulates the problem as an interval multi objective optimization model, where path length and safety risk are optimized simultaneously, and the safety risk is represented as an interval based on the robot's relative distance to hazardous sources. On this basis, a decomposition based interval multi objective evolutionary algorithm with dynamic neighborhood and adaptive weight adjustment, called IMOEA/D-DAW, is proposed. The algorithm introduces an interval sparsity function to evaluate the distribution of solutions on the Pareto front, and then uses this information to adaptively adjust weight vectors and neighborhood sizes according to subproblem sparsity and evolutionary progress. In this way, both convergence and diversity are improved during the search process. Experimental comparisons with representative algorithms demonstrate that the proposed method produces better path planning results in uncertain chemical plant environments, with advantages in both convergence and solution diversity.