Case synthesizing expert system for network-based project planning

Abstract

A construction manager uses a project plan network which is represented by the form of PERT/CPM to analyze resource requirement, whole duration, and so on. Generating a large and complex bridge construction plan requires much knowledge and high expertise of domain experts. Since there has been no standard and the past network itself contains planning knowledge, we have used CBR to generate bridge construction plan based on network. It is difficult to generate bridge construction plan by single case adaptation because of the wide variation of bridge plan. So we developed multiple sub-cases synthesizing methodology to solve the problem of variation. For resolving this, we developed a case synthesizing architecture. This paper is reviewing much literature and classifing them with three criteria: Ability to be decomposed, Sensitivity to change on network, Possibility to find similar case. For generating new project network, user input basic information into system. Then, system do analyze and decompose the problem, partial search and retrieve, adapt to fit current problem, and synthesize. Adaptation in this domain is generate and/or delete of activities and those precedence relationships and determine the duration for activities. Synthesizing in this domain is resource constrained scheduling problem. If the result is not satisfiable, re-do adaptation and synthesizing with revised constraints. We developed an application to use in real world by this architecture. This study suggests the criteria for selecting an adaptation methodology by the characteristics of the case. It presents a Case Synthesis Methodology for network based project planning. We develop an architecture and an application by using the Case Synthesis Methodology. The application has been used on real world. This architecture can be a frame for pre analysis. It generates a Reliable project plan, enables knowledge collection. And it also expected to bring business engineering effect, standardization ...