Planning and scheduling highway construction using GIS and dynamic programming

This study presents a model to plan, schedule and control the construction of highway projects.

A two-state-variable, N-stage dynamic programming procedure, in conjunction with a set of heuristic rules, is capable of generating schedules to meet a specified budget or duration restraints, and can select the optimum expediting technique to reduce project duration. The proposed model incorporates both typical (equal quantity of work in all units) and non-typical (varying quantities along highway length) repetitive activities in the optimization procedure. The proposed model accounts for the impact of practical factors, such as transverse obstructions, on crew assignment and project duration and cost. The impact of inclement weather on crew productivity is accounted for, as well as the beneficial effects of the learning curve. A relational database model is developed to store: (1) local weather conditions, (2) swell and shrinkage factors for various soil types, and (3) the available resource pool. The model supports tracking and controlling project progress utilising the earned value concept. A flexible and practical methodology is proposed to forecast project time and cost at completion, enabling the assignment of a higher weight to recent crew performance.

A prototype software based on the model is developed, coded in Visual C++ (version 6.0), employing object oriented programming. ArcView^® version 3.1 is employed as the GIS engine, and the 3-D data analysis is carried out employing ArcView^® 3D Analyst. (Abstract shortened by UMI.)