Modeling and Analysis of Dynamic Assignment in Scale-Free Networks: An Agent-based Simulation Exercise

An Agent-based Simulation Exercise Yu Teng, Nan Kong* Weldon School of Biomedical Engineering, Purdue University 206 S. Martin Jischke Dr., West Lafayette, IN 47906 ABSTRACT This paper addresses stochastic dynamic assignment of perishable goods in a scale-free network. We develop an agent-based simulation model in which the supplies and demands for a single perishable commodity are generated at each node (agent) in the network. We match supplies and demands with several pre-specified assignment rules and given various commodity decay durations. A utility is associated with each assignment based on the spatial proximity between the supplier and the recipient. We assess various rules in terms of overall network utility. Our computational results show that 1) it is beneficial to use hybrid assignment rules that integrate the spirit of system-wide assignment and that of local preferred assignment; 2) it may be beneficial to allow agent behavior changes with certain degree of autonomy.