Primordial abilities in artificially intelligent systems for civil engineering design
Adewale Abiola Oki · Spiral (Imperial College London) · 1990
This thesis addresses primordial deficiencies associated with existing knowledge-based systems for civil engineering design.It is contended that such systems lack innate abilities which human beings are adept at exploiting, and which are therefore desirable in artificially intelligent systems for design.Three abilities are explored: reasoning using previous experience, intelligent utilization of knowledge, and reasoning using pictorial information.The investigation focusses on employing artificial intelligence techniques for devising procedures by which these abilities may be realized, engendering a more effective and representative computational model for civil engineering design simulation.Appropriate frameworks are proposed and subsequently implemented in a logic/procedural programming environment as a precursory step -a positive consideration of heretofore neglected directions.However, this is carried out with an underlying implication of synergistic behaviour.The abilities (amongst others) are then colligated with a model for structural design, by means of a set theoretic elucidation, whereby the model acquires a more general character.A prototypal knowledge-based system for the preliminary design of structural timber roofs is further developed with the aid of the model.The thesis points up the argument that the need for superlative research on fundamentals is indubitable if such systems are to play a quintessential role in design.