The use of an appropriate Project Delivery System (PDS) can significantly increase the project efficiency and success rate. Designing a PDS however is a difficult task because: (1) decisions are made early in the project planning phase when only limited and imprecise project information is available; (2) decisions involve choosing between a variety of PDSs and consideration of multiple factors (e.g., project characteristics and external environment factors); (3) decisions are made in a multi-project environment with complex project dependencies. Several methods (e.g., guidance, multi-attribute utility analysis, and analytical hierarchy process) have been proposed to help a decision-maker choose a PDS. Nonetheless, project dependencies and possible time/cost trade-offs are frequently ignored in these methods even though they can affect project performance. In addition, these methods provide very little, if any, opportunity to the decision-maker for "designing" a PDS for their needs. This research seeks to overcome these limitations.

This research develops a decision-support framework for designing and choosing a PDS in a multi-project environment. The framework was developed for the specific circumstances of one public agency, which was required by regulations to choose from design-bid-build (DBB), design-build (DB), or construction management at risk (CMR), with limited flexibility to adapt these contractual structures to the specific circumstances of individual projects. That does not materially impact this research, the focus of which is incorporating time/cost tradeoffs and interdependencies in a multi-project environment into the design of PDSs. This framework helps a decision-maker evaluate alternative PDSs (in this case, specifically the DBB, DB, CMR alternatives) with respect to two groups of criteria: performance and general PDS criteria. It also integrates a Net-Present-Value-based (NPV-based) method for objectively determining time/cost trade-off rules and a procedure for systematically evaluating effects of project dependencies on project schedule and cost.

The case study results show that the framework is beneficial to a PDS decision-maker in several ways. First, it guides a decision-maker, step-by-step, in evaluating alternative PDSs. With the information collected and analyzed in the framework, the decision-maker can better justify his/her decisions to others. Furthermore, it helps the decision-maker consider project dependencies and possible time/cost trade-offs in estimating project timelines and cost distributions in different PDSs. Such consideration can facilitate not only a more realistic project planning but also a more informed PDS decision-making. The proposed framework can also provide the decision-maker opportunities to explore "what if" scenarios in his/her PDS decision-making process. Finally, but not least, the framework also allows the decision-maker to identify the areas where a PDS may perform poorly, and thereby develop proactive management strategies.

The proposed framework has one problem: its application is more time-consuming than some existing PDS selection methods (e.g., a weighted score approach). This longer application time is partially because of the unfamiliarity of decision-makers with the framework, and hence can be improved through proper training. The needs to estimate project timelines and cost distributions in different PDSs and to evaluate effects of project dependencies also contribute to this delay. To ensure a more effective framework implementation in the future, public agencies are suggested to: (1) establish proper documentation and knowledge management systems, (2) develop an effective communication structure, and (3) to provide proper training to enhance the project managers' competence. A strong support from upper-level management is also critical to a successful framework implementation.