Diffusion of Innovation Simulation Using an Evolutionary Algorithm

The diffusion of innovation theory aims to explain how new ideas and practices are disseminated among social system members. A significant number of the existing models is based on the use of parameters which determine the process of innovation adoption, and rely on simple mathematical functions centered in the observation and description of diffusion patterns. These models enable a more explicit diffusion process study, but their use involves the estimation of diffusion coefficients, usually obtained from historical data or chronological series. This raises some application problems in contexts where there is no data or the data is insufficient. This paper proposes the use of evolutionary computation as an alternative approach for the simulation of innovation diffusion within organizations. To overcome some of the problems inherent to existing models an evolutionary algorithm is proposed based on a probabilistic approach. The results of the simulations that were done to validate the algorithm revealed to be very promissing in this context.

Simulation experiment results are presented that reveals a very promising approach of the proposed model.