Abstract
A decision tool that can accurately predict the market diffusion of a new product can be helpful in the establishment of product design. This study simulates product diffusion through multi-agent simulation (MAS) and determines product specifications that can maximise future sales by adapting ordinal optimisation. MAS constructs a virtual market of agents (consumers) and predicts the sales volume by simulating the product selection process of individual agents. By interacting with the MAS, the ordinal optimisation rapidly determines the product specifications of maximum sales with few computations. An empirical experiment conducted on the Korean smartphone market yielded interesting results, i.e., it is desirable for companies with high brand values not to offer low-performance products but to make additional efforts to increase satisfaction with hardware specifications, while a high-end policy of product quality and a low-price sales strategy may be more beneficial for increasing sales volumes for companies with low brand values.