Cooperation and competition in inventory games (Q1423701)

In an inventory game, \(n\) agents agree to combine their orders, so as to save on pre-order costs. In an optimization model that takes into account the ordering costs and the storage costs, it can be proven that when each agent \(i\) submitted, on average, \(m_i\) orders per time unit, then the coalition of agents would need to make only \(m=\sqrt{\sum m_i^2}<\sum m_i\) orders. In this paper, the authors prove that the same estimate for \(m\) holds in a more realistic model in which we also take into account possible losses due to shortages (while waiting for the order to be fulfilled). For both models, the next natural question is: what is the fair way to distribute the ordering costs between the participating ordering agents. Previously, only a semi-heuristic cost allocation rule was known. The authors provide reasonable axioms for the cost allocation and prove that these axioms uniquely determine this allocation rule.