Parametric Constraint Management in Planning and Scheduling: Computational Basis

Abstract

Critical path methods (CPM) have long been used as the computational basis for computer scheduling systems. The basic computational methodology in CPM proved to be inadequate to represent and integrate more problems in project management. The main reason is that the methodology is built on a computational procedure that is specific to the underlying network representation. Various knowledge-based techniques have been used to provide a richer computational methodology. The main goal of this paper is to describe the computational basis of a constraint-management methodology as applied to planning and scheduling. Each problem may be represented by interrelating its parameters using construction duration and precedence knowledge. The parameter relationships may be viewed as mathematical constraints. The paper provides a description of the computational methodology including the constraint modeling process, the graph representation of constraints, and the evaluation of constraint networks. The methodology may result in multiple solutions based on the nature of graph representation. Therefore, the paper proceeds to prove that only the acceptable computational solution represents the appropriate project plan and schedule. The paper concludes by commenting on the utility of the methodology and providing directions for future research.