New product development is a high-risk decision-making problem in which similar products compete with each other to expand their market shares. Brand-level diffusion predictions can help product design managers to analyse how product attribute specifications impact total market shares, which can, in turn, aid managers in choosing the designs that yield maximum profits. In this paper, we develop a product attribute design method in which an artificial market consisting of consumer agents in an interaction network is created to simulate the diffusion process of products, and a genetic algorithm is integrated with the artificial market to support the product design decision-making process. The contribution of this research is that the predicted market response to product alternatives is incorporated into the product design optimisation. Two empirical experiments were conducted on the Korean laptop computer market to demonstrate the potential of this integrated method. Preliminary experiment showed that our prediction diffusion curves had an average error of 3.04%. In the primary experiment, five designs were recommended, and a comparison with the 31 best-selling laptop computers resulted in an average error of less than 8% when the "Price" attribute was excluded.
All Science Journal Classification (ASJC) codes
- Modelling and Simulation
- Computer Science Applications