Various types of deep foundation result in difficulty when excavating and supporting in soft soil strata during construction. An unpleasant situation is that too much project cost have to be invested in foundation engineering, among which quite a lot is dedicated to supporting excavation measures. Choosing a proper supporting system with low-cost, without much detrimental influence to environment during construction, become main concerns of an engineer. Several techniques based on computer methods are developed for this goal. Phase I of this investigation places emphasis on scheme planning where there exists complexity of soil conditions, environmental factors, variable types of retaining structure and so on. A new technology for providing information to scheme design is proposed in the light of fuzzy decision processes. To make use of previous knowledge, a Geographical Information System is also developed on the basis of previous foundation records. In the study of Phase II the method to build an expert system for a retaining structural design is systematically discussed. It organizes the knowledge about a excavating project from a rational point of view. Phase III concentrates on optimization of a series of suitable schemes. Comparison with in-situ measures shows an applicable and efficient procedure.
Constructions of tall buildings or underground engineering cannot avoid deep excavation. This situation becomes serious in protecting the safety of existing surrounding facilities, especially when a site is located in an urban area. Since the first application of fuzzy theory presented by Bellman and Zadeh (1970), the fuzzy approach has seen many developments in other disciplines, such as mathematical programming problems (Zimmermann, 1985), fuzzy optimum design of structure (Wang and Wang, 1985) and mechanical system (Rao, 1987).