Particle swarm optimization for optimal product line design

Designing optimal products is one of the most critical activities for a firm to stay competitive. Except for genetic algorithms, previous approaches that solve the optimal product line design problem provide the decision maker with a single best solution. Furthermore, they assume a static market, in which the incumbent firms will not respond to the entrance of a new player. In this paper we apply a new population-based algorithm called particle swarm optimization to the problem and employ a Monte Carlo simulation to compare its performance to that of genetic algorithms. The results indicate that the proposed particle swarm optimization algorithm constitutes an attractive alternative for solving the optimal product line design problem because its performance is comparable to that of genetic algorithms concerning the best solution found while it outperforms genetic algorithms regarding the diversity of the final set of provided solutions. Furthermore, we use concepts from game theory to illustrate how the algorithm can be extended to incorporate retaliatory actions from competitors. The dynamic approach is illustrated through a real-world case in which a firm intends to enter the Greek retail milk market. While employing highly simplifying assumptions, the incorporation of the Nash equilibrium concept provides useful insights, such as the attribute levels that may be resistant to competitive reactions and the incumbent firms that will benefit most in the long term.