A Hybrid Multi-Agent/Spatial Interaction Model System for Petrol Price Setting

This paper outlines the findings of ongoing research on the application of multi-agent simulations (MAS) to modelling complex locally interacting dynamic systems, in particular the petrol price market. The development of an agent based model for petrol prices is presented. Failings within this model are discussed and an alternative strategy for controlling the price of each petrol station based on profit maximisation is considered. A spatial interaction model was used to estimate the sales and linked to the agent system to create a hybrid model. To evaluate how effective this hybrid model was, a comparison was made with an existing data set of real petrol prices collected over a two month period. This was achieved both statistically and visually with the use of a Geographical Information System (GIS). Experimentation revealed that the hybrid model outperformed the agent model. Investigation into the behaviour of the system (how prices diffuse spatially) was undertaken by means of idealised simulations.