Abstract

The beer distribution game has been extended into agent-based model applicable for complex networks to investigate how internal decision making in multiple interconnected stocks under disruptions can influence the overall system performance. Two most common topologies have been taken into account: random and scale-free. These have been used as a structural base for the agent-based simulation, where a node represents a single stock and a link represents the material flow. Disrupted networks have safety stock and contingency rerouting strategies applied as recovery actions. Results show that scale-free networks exhibit higher robustness, they generate lower costs, have higher fill rate and recover better from disruptions. Safety stock proves to be an effective risk mitigation strategy applied in a complex supply network, since it increases fill rates and might decrease the cost depending on the risk profile. Contingency rerouting proves to be an effective mitigation strategy only for long disruptions, where it decreases costs and increases fill rate.