Front-end planning of construction sites can have a significant impact on the safety and efficiency of site operations and/or the cash flow associated with resource management. A well-planned site can:​ (i) minimize travel time;​ (ii) decrease time and effort spent on material handling;​ (iii) increase productivity;​ and (iv) improve safety, and hence decrease construction cost and time. Despite its importance, it is often ignored in the planning phase of construction projects. In practice, space allocation on construction sites is typically carried out on a first-come-first-serve basis, which could result in chaotic sites and productivity losses.

This thesis presents an interactive CAD-based model, designed to support site layout planning and to provide a framework that meets the versatility needed in actual construction practice. The developed model performs its task at two levels:​ site representation, and site space analysis and allocation. The site representation is carried out using object-based concepts in an open architecture. The model offers three tiers of objects:​ (i) site objects, (ii) construction objects, and (iii) constraint objects. This classification assists in formalizing the representation needed for modeling construction sites. A formal structure for each tier of objects is proposed to facilitate the creation of new objects and reuse of domain knowledge. The space analysis and allocation is performed through a geometric approach, inspired by human reasoning, in search for the optimum or near-optimum location for temporary facilities.

The model has a number of interesting features:​ (i) it provides a flexible support of a wide range of objects for site planner. This permits the set up of different construction projects via selection of objects stored in their respective libraries;​ (ii) the model has a built-in feedback to support the development of new objects and/or updating existing ones. This allows for the gradual expansion and enrichment of the supporting libraries;​ (iii) the geometric space analysis is visualized graphically and provides a range of ranked near optimum solutions. This feature facilitates user's comprehension of, and interaction in the layout process;​ and (iv) the system allows site planners to define search criteria based on their knowledge and expertise. It also allows for generation of a number of what-if scenarios, utilizing different rules and comparing the final layout results.

A prototype of the developed model called CASL is implemented using Visual Basic for Applications (VBA) in AutoCAD environment. AutoCAD is dynamically linked to the relational database of Microsoft Access, which hosts the object libraries. To validate and illustrate the functionality and usability of the developed model, two case studies from the literature were analyzed using CASL. Further, a case study from industry was developed to illustrate the flexibility and other features of CASL that are not available in other models. Comparisons between the results generated by the developed model and those provided by others indicate that CASL is capable of generating layouts that better satisfy the constraints defined by planners. As such, CASL provides users with a design support tool, capable of utilizing their knowledge in generating an efficient site layout, which can aid to avoid cost overruns, schedule delays, and unsafe working conditions.