Using construction method models to generate four-dimensional production models

Construction planning is an important task on all construction projects; it establishes a link between the facility design and the process with which it will be created. Construction planners generate plans for a variety of reasons, e.g., to study the feasibility of a project from a temporal perspective, to evaluate the constructibility of a design alternative, or to develop a detailed, tactical plan for the day-to-day operations of a site. Regardless of a plan's purpose, it is desirable to generate realistic plans at the appropriate level of detail as quickly and cheaply as possible. Today's construction planning process, however, is largely a manual and time-consuming process that does not support the rapid generation of plans. The translation of design information to a construction plan requires manual interpretation of data and knowledge in the planner' s mind. Construction plans generated with this "informal" construction planning knowledge are difficult to maintain because the plan does not represent why activities and their sequencing links exist. The main problem is that construction planning information exists in many different forms and places (in planners' minds, on paper, in computer files) and is neither integrated nor easily accessible. In this dissertation, I address this problem and propose an integrated planning process and computer-interpretable models that support activity generation and sequencing. I formalize a 4D production model for the Architecture, Engineering, and Construction (AEC) industry. The 4D production model consists of all activities needed to construct a facility. I formalize the representation of an activity with the following fundamental construction entities: components, actions, resources, and sequencing constraints (abbreviated as a <CARS> tuple). With these entities planners can model who does what, when and where at multiple levels of detail. For all but the smallest construction projects, it is impractical to generate a 4D production model manually. Therefore, this dissertation also formalizes a hierarchical method-driven planning process. I have formalized a Construction Method Model Template (CMMT) that allows planners to model the activities required for a particular method. Planners define each activity in a CMMT with <CARS> entities. The Construction Method Modeler (CMM) prototype system implements the hierarchical method-driven planning process. With CMM, planners can rapidly generate construction plans and explore the impact of changes in material and labor availability and study the effect of design changes. The 4D production model also allows them to communicate a plan as a visual 4D model which helps to identify interference with other crews. In summary, CMM helps planners to understand the impact of overall project decisions (e.g., a change in completion date) on day-to-day resource allocation and assists in showing the effect of day-to-day progress on overall project progress.