An A.i. Approach to Automating the Recognition and Evaluation of Logic Classes in Construction Scheduling

Construction schedules exist in many forms to serve many purposes. They are prepared by various means by different individuals either manually or aided by computer software. Work which is underway to automate the production of schedules is now being supplemented by the development of an evaluation package for general usage . Outlined in this paper are aspects of a research project which uses artificial intelligence techniques for the purpose of automating the evaluation of construction schedules. In particular the early stages of the project, which recognise and utilise the logic in schedules, are detailed. Different classes of logic are identified and a general activity logic model is introduced. Techniques for interpreting activity names and for building and using knowledge bases are explained and illustrated with simple examples. The application of the early work to the objective of automated schedule evaluation is discussed. INTRODUCTION In recent years there has been some interest in automating the production of construction schedules and a number of approaches using artificial intelligence techniques have emerged for tackling the problems (1,2). Any developed system emerging from these approaches should be capable of self -evaluation for users to have increased confidence in the results. At present schedules are produced by individuals (or small teams) working from their own base of knowledge using algorithmic computer software for the benefits of speed and presentation. Evaluation, if it occurs at all, is likely to be little more than a brief check by a colleague or a rudimentary appraisal by a client's representative. Checking may only be for gross errors because of the considerable diversity in end results which different individuals could achieve. The end results could also be presented in many forms such as * a barchart * a network * a space time diagram. Schedules may also be produced to meet the needs of different groups such as * the client * the consultant * the contractor. They may further contain information represented to different levels of detail such as