A Multiobjective Optimization Model for Packing Proposals in Large-Scale R&D Project Review

Abstract

To solve the packing proposals problem in large-scale R&D project review, a-nonlinear 0-1 integer programming model with multiobjective is presented. The first objective is to minimize the quantity difference between the number of proposals for each package and the average number of proposals that peer reviewers would like to review. The second objective is to minimize the probability that proposals belong to the same organization be packed into one package. A weighted sum method is applied to treat multiple objectives. Then, a genetic algorithm with sequence encoding and PMX crossover is developed to solve this model. Finally, a numerical experiment designed to simulate the proposal packing process of NSFC is described to illustrate the use of the proposed method