A GENETIC ALGORITHM FOR RISK-BASED PARTNER SELECTION PROBLEM IN NEW PRODUCT DEVELOPMENT

摘要

The problem of risk-based partner selection in new product development is described and the total development risk is minimized by choosing optimal partners for each activity. Because risk coefficients are expressed in the form of interval numbers, the model is a nonlinear programming with interval coefficients in its objective function. In order to solve the problem, we give the definition of order relation and convert the original problem to equivalent biobjective problem. A genetic algorithm with adaptive evaluation function is applied to find the set of Pareto or near Pareto solutions of the biobjective problem. Simulation results show that the algorithm is efficient and the model has potential to practical applications.