This paper presents a new class of two-stage random fuzzy programming with recourse (RFPR) problems. Since the RFPR problem usually includes random fuzzy parameters with infinite supports, it is inherently an infinite dimensional optimization problem that can rarely be solved directly by the conventional optimization algorithms. To overcome this difficulty, this paper developed an approximation method for the original RFPR problem, and turn it into a finite-dimensional one. We also establish a convergence relation between the objective values of the original problem and its approximating problem. To solve a general RFPR problem, we design a hybrid algorithm by integrating the approximation method, neural network (NN) and particle swarm optimization (PSO) algorithm. Finally, one numerical example is presented to demonstrate the effectiveness of the designed algorithm.
展开▼
