Despite the importance of the management of inventory to industrial scheduling applications, there has been little research that has directly addressed the production, consumption, and storage of inventory as part of a scheduling problem. In this paper, we extend the job shop scheduling problem model by directly representing inventory, inventory storage constraints, and inventory production and consumption in a constraint-directed scheduling framework. Inventory scheduling is then used to investigate the formulation of heuristic commitment techniques based on the understanding and the exploitation of problem structure. An estimation of probability of breakage for resource and inventory constraints is presented together with a number of heuristic commitment techniques with varying dependence on the estimate of constraint criticality. It is empirically demonstrated that the heuristic commitment technique that exploits dynamic constraint criticality achieves superior overall search performance. Copyright © 2002 John Wiley & Sons, Ltd.