A hybrid particle swarm optimization algorithm for satisficing data envelopment analysis under fuzzy chance constraints

This paper presents a new satisficing data envelopment analysis (DEA) model with credibility criterion, in which the inputs and outputs are assumed to be characterized by fuzzy variables with known membership functions. When the inputs and outputs are mutually independent trapezoidal fuzzy variables, we turn the proposed satisficing DEA model into its deterministic equivalent programming problem. For general fuzzy input and output variables, we design a hybrid particle swarm optimization (PSO) algorithm by integrating approximation method, neural network (NN) and PSO algorithm to solve the proposed DEA model, in which the approximation method is used to compute the credibility functions, NN is used to approximate the credibility functions, and PSO is used to find the optimal solution of the proposed DEA problem. Furthermore, the sensitivity analysis of the proposed model is discussed. Finally, we perform a number of numerical experiments to demonstrate the effectiveness of the hybrid PSO algorithm. The computational results show that the designed hybrid PSO algorithm outperforms the hybrid genetic algorithm (GA) in terms of runtime and solution quality.