Publications / 1996 Proceedings of the 13th ISARC, Tokyo, Japan
Self-programming, intelligent, flexible machine systems offer a desirable alternative to high-volume, hard automation solutions in the complex, ill-structured and dynamic field of construction. A theory of construction is developed which shows that the nature of construction requires the capability for problem solving. Highly complex, dynamic problems require a high degree of intelligence, or problem-solving flux, for solutions to be found. An opportunity exists to apply machine intelligence capable of autonomously attacking construction work. Economic viability underpins the design problem, and the specifications developed in this paper for an effective solution show that self- programming systems are preferable for the low-repetition environment of construction. An architecture for a self-programming system is developed, showing application and multi-agent potential, and highlighting the need for such systems to be developed.