Product portfolio planning: a metaheuristic-based simulated annealing algorithm

Product portfolio diversity is prominent for customers, but critical for manufacturers. From the manufacturers’ perspective, diversity must be maintained at a level where engineering costs do not exceed the acquired advantages of increased market share. Therefore, in this paper we consider the problem of product portfolio planning to simultaneously maximise market share and minimise manufacturing engineering costs. Since this problem belongs to the NP-hard class of problems, exact algorithms are incapable of rendering an optimal solution. Therefore, we used metaheuristic-based simulated annealing to deal with the problem. Our proposed algorithm consists of two parts, i.e. a random search and a predetermined rule to generate the next possible neighbours (product portfolio). In order to have a robust algorithm, we calibrated different levels of our problem's parameters using the Taguchi method. This method picks the best levels of different parameters, conducting the least possible experiments. To evaluate the performance of our proposed algorithm, we compared it with a strong algorithm – the genetic algorithm. We used this comparison as the basis of our research. The obtained computational results clearly demonstrate the efficiency and effectiveness of our proposed algorithm.