Constraint knowledge for construction scheduling

The capabilities of network diagramming techniques (NDT) are restricted by limitations inherent in their representation of schedule constraints (typically expressed through precedent relationships between activities). This assertion suggests the selection of a richer representation as a departure point for extending the utility of planning and scheduling techniques. The authors suggest that such a representation is provided by a system which employs a general model of a project´s `status´ and which allows schedule constraints to be expressed as rules which refer to this status. This system offers the advantages of allowing the precedence of activities to be based on more than just the completion of other activities. It also provides an efficient knowledge-based approach to scheduling that can express the reasoning underlying scheduling actions which can be employed by future artificial intelligence (AI) planning. The authors discuss how constraints are represented and describes what types of constraints can be represented in both NDT and in the proposed system-called A Construction Planner (ACP). This is followed by a comparison of the schedule generation algorithms used in NDT and in ACP