We consider the problem of rational, self-interested, economic agents who must negotiate with each other in order to carry out their plans. Customer agents express their plans in the form of task networks with temporal and precedence constraints. The market runs a combinatorial reverse auction, in which supplier agents submit bids specifying prices for combinations of tasks, along with time windows and duration data that the customer may use to compose a work schedule. The presence of temporal and precedence constraints among the items at auction requires extensions to the standard winner-determination procedures for combinatorial auctions, and the use of the enhanced winnerdetermination procedure within the context of a real-time negotiation requires that we predict its runtime when planning the negotiation process. We address two specific issues related to this problem. The first is the need for a market infrastructure to support decision processes. We propose a set of requirements for a market that can support this type of negotiation, and describe an architecture that can meet these requirements. We also describe the high-level design of an agent that can act as a customer in this environment, and discuss the decision behaviors such an agent must implement to maximize its utility. The second issue we consider is the determination of auction winners. We explore and characterize a winner determination method, which is an extension of the bidtree-based Iterative-Deepening A* (IDA*) formulation proposed by Sandholm.