Agent-based simulations coupled with an analysis of the flow of electric power are carried out to examine the influence of the social policy of the government and the neighboring communication between customers on the adoption of distributed rooftop photovoltaic electrical power generators (PVs). How the relationships between the social policy and the possibility of a reverse current restriction give rise to the collective behavior of autonomous individuals, and how the end customers interact and form relationships with its environment are described. Strong intervention by the government in the areas near a main high-voltage power distribution transformer, where the possibility of a reverse current restriction is relatively low, drives the greatest adoption of the PV system. The near areas are primarily occupied by customers with only a PV to improve the diffusion rate of PVs via the self-organization by the communication between customers. It also lead in a decrease in the need for compensation devices, which in turn minimizes the social cost. Growth in the number of PVs in areas far from the transformer is assisted by the installation of batteries as compensation for the lost opportunity due to restrictions in those areas on reverse power currents. Therefore, excessive intervention by the government in the far areas results in an increase in the social cost of managing reverse currents. (C) 2014 Published by Elsevier Ltd.