Combinatorial Trading Mechanism for Task Allocation

摘要

We describe a task reallocation problem among agents, and present on a market-based solution procedure. A new bidding form, which allows agents to sell and buy tasks simultaneously, is proposed. Based on the bidding form and the concept of feasible trading matching, an iterative combinatorial trading mechanism is devised to extend the one-side combinatorial auction mechanism. We show that this mechanism can produce optimal task allocations. An algorithm for feasible trading matching with maximal trading surplus is also presented. Finally, we extend the trading mechanism to deal with task reallocation and team adjustment simultaneously.