Article ID: | iaor20082319 |
Country: | United Kingdom |
Volume: | 34 |
Issue: | 6 |
Start Page Number: | 1777 |
End Page Number: | 1799 |
Publication Date: | Jun 2007 |
Journal: | Computers and Operations Research |
Authors: | Wang Yi, Jiao Jianxin (Roger), Zhang Yiyang |
Keywords: | heuristics: genetic algorithms |
Product portfolio planning has been recognized as a critical decision facing all companies across industries. It aims at the selection of a near-optimal mix of products and attribute levels to offer in the target market. It constitutes a combinatorial optimization problem that is deemed to be NP-hard in nature. Conventional enumeration-based optimization techniques become inhibitive given that the number of possible combinations may be enormous. Genetic algorithms have been proven to excel in solving combinatorial optimization problems. This paper develops a heuristic genetic algorithm for solving the product portfolio planning problem more effectively. A generic encoding scheme is introduced to synchronize product portfolio generation and selection coherently. The fitness function is established based on a shared surplus measure leveraging both the customer and engineering concerns. An unbalanced index is proposed to model the elitism of product portfolio solutions.