Ghost: Project Network Generator

Abstract

Network scheduling systems have flourished in the construction industry since their inception in the 1950s. Today, there is a plethora of project management programs that incorporate sophisticated techniques such as resource leveling, time‐cost tradeoff analysis, etc. Despite the fact that all network‐based techniques depend on the existence of a sound project network, very little attention has been given to the development of algorithms for generating the network. This is not surprising because generating a network is a complex heuristic process, for which computationally efficient algorithms do not exist. It is only recently that developments in artificial intelligence have made it possible to address the problem. This paper describes a prototype knowledge‐based project network generator (GHOST). GHOST takes as input a set of activities and produces as output a schedule by setting up precedents among the activities. The knowledge base is made up of several knowledge sources known as critics. The critics contain knowledge about basic physics, construction norms, redundancy in networks, etc. A methodology and an extended example are provided in the paper.