Mario Sillus

Abstract

Storage resources represent material stocks or liquid funds that are depleted and replenished at the occurrence of certain events during the execution of a project. Both renewable and nonrenewable resources are special cases of storage resources. Project scheduling subject to storage-resource constraints and generalized precedence relations consists in sequencing the project events in such a way that the inventory levels of the storage resources evolve within given bounds and prescribed minimum and maximum time lags between the events are met. Such a scheduling problem arises, for example, in production planning when material-availability or storage-capacity constraints for intermediate products must be observed or in large-scale projects where cash in- and outflows need to be coordinated to avoid liquidity crunches. In presence of storage resources, even the feasibility variant of the scheduling problem is NP-hard.

In literature, different enumeration schemes for project scheduling problems with storage resources and regular objective functions like the project duration have been devised. We develop a new time-based schedule-generation scheme that is intended to serve as a building block for metaheuristic schedule-improvement procedures. The method decodes an event list into a feasible schedule by iteratively resolving resource conflicts. An inventory shortfall or excess is stepwise reduced by increasing or decreasing lower bounds on the occurrence times of appropriate events. The generation scheme is complete and equipped with some pinch of randomness and unlimited computation time it is guaranteed to provide a feasible schedule in a single pass.