| Abstract: | Periodic overhaul planning for boilers and turbines in electric power plants has been done based on expert experience in order to satisfy various constraints that are hard to formalize, such as legal regulations, the overhaul interval restriction between units, the adjustment strategy in the case of including special events, the daily electric power requirement, and so forth. Although computer processing using the knowledge base approach has been tried, it does not lead to good results because of excessive hardware requirements, low performance, and no assurance of optimality. A two‐step approach is proposed based on the categorization of constraints that are hard to formalize into “strong constraints” and “weak constraints.” It replaces the nonlinear integer programming problem corresponding to “strong constraints” by a combination of integer programming and combinatorial analysis. Consequently, it can lead to optimal solution in terms of fuzzy satisfaction measures within an allowable time with a PC. © 1999 Scripta Technica, Electr Eng Jpn, 127(2): 24–31, 1999 |
