Methodology for modelling a project management control environment

Control must be interpreted in its widest sense if its substantial body of knowledge and those who have helped to establish this are to remain gainfully employed in the solution of realworld engineering problems. Many of these, in an age properly described as large-scale, are found in the field of project management. Herein lie control problems of the most challenging nature, to the extent that researchers are proposing an essentially different control strategy than has hitherto been realisable: that of the intelligent knowledgebased system (IKBS). The paper aims to formalise the expertise of a project manager, regarding the two specific problems of bidding control and expedition control, and hence to develop a suite of user-friendly computer-based tools embodying this knowledge. The paper proposes an original and powerful methodology for modelling projects. This is based on an information system represented by a type of associative network. The authors have made an important modelling decision to classify all information into distinct planes in terms of project specifics, project environments and project generalities. Project knowledge is seen as potential control action, and this is modelled as a reverse-chaining inference system attached to slots of frames. Reasoning with uncertainty is provided for as is an explanation facility. The means of representing intelligence is unique in that the inferencing, instead of operating from a central database containing evidence and hypotheses, works in conjunction with the information system which maintains both. Two IKBSs are described which are aimed at the problems of bid proposal and project expedition, respectively. These systems have emerged from a background of extensive consultations with industries intimately concerned with project management.