Pareto Process Optimization of Product Development Project Using Bi-objective Hybrid Genetic Algorithm


Overlapping and iteration between development activities are the main reasons to cause complexity in product development (PD) process. Overlapping may not only reduce duration of a project but also create rework risk, while iteration increases the project duration and cost. In order to balance the duration and cost, this article presents four types of time models from the angle of time overlapping and activities dependent relationships based on Collaboration Degree Design Structure Matrix (CD-DSM) and builds the cost model considering the negation cost. On basis of the formulated model, a hybridization of the Pareto genetic algorithm (PGA) and variable neighborhood search (VNS) algorithm is proposed to solve the bi-objective process optimization problem of PD project for reducing the project duration and cost. The VNS strategy is implemented after the genetic operation of crossover and mutation to improve the exploitation ability of the algorithm. And then, an industrial example, a LED module PD project in an optoelectronic enterprise, is provided to illustrate the utility of the proposed approach. The optimization model minimizes the project duration and cost associated with overlapping and iteration and yields a Pareto optimal solution of project activity sequence for project managers to make decision following different business purposes. The simulation results of two different problems show that the proposed approach has a good convergence and robustness.