ABSTRACT
Knowledge-based problem solvers traditionally merge knowledge about a domain with several heuristics in an effort to confront novel problem situations intelligently. Solutions to many real-world problems involve searching for an optimum solution satisfying several constraints. This chapter presents an intelligent hybrid system, called WATTS (WAsTewater Treatment System), that can acquire knowledge from examples and use that knowledge to obtain optimal treatment trains for wastewater. Two different methods, heuristic search and neural network approaches, are presented for generating optimal solutions. Also presented is an on-line learning capability using a case-based reasoner that can store old solutions and use them for solving new problems in both approaches.
