Innovation diffusion in a heterogeneous population

One of the assumptions made in modelling innovation diffusion is that of complete mixing of prior and potential adopters. In the present paper we give a deterministic model for diffusion of innovation in a population for which this assumption is inappropriate. The population of interest is divided into two spatially separated groups. Diffusion in the first group near the single centre of innovation is assumed to follow the logistic model. A modified model is proposed for the second group. Individuals in this group are assumed to receive information from previous adopters in both the groups. The model is fitted to data on diffusion of crossbred goats in villages around Narayangaon, a town about 70 kilometers from Pune in Western India.