Knowledge-based expert systems have been developed for many civil engineering applications including bridge deck condition assessment, selection of optimal strategies for bridge deck rehabilitation, and traffic signal setting. Such expert systems incorporate both heuristic knowledge and algorithmic approaches to problem solving. Identification of bridge-painting strategies is perfectly suited to such an approach. Bridge-painting decisions are based on measurement of condition; qualitative assessment of deterioration; and heuristics describing the incompatibilities among different types of steel, paint, and surface preparation. Further, uncertainty plays a crucial rolebecause surface treatment, paint application, and bridge condition are nonuniform. Optimization or current approaches to decision making are unable to effectively include all of these variables. A prototype system, bridge piars (paint identification and ranking system), constructed using an expert system building program is based on a decision network. The system allows the user to establish the facilitycondition, evaluate the need for bridge painting, identify appropriate painting strategies, and cost the strategies. The system and itsoperation are described, and several areas for research to extend and enhance the system are identified. This paper appeared in transportation research record no. 1145, Expert systems for transportation applications. For covering abstract see IRRD no 817754.
Abstract