Programming of pavement maintenance is a difficult optimization problem that suffers from combinatorial explosion due to the very large number of possible combinations involved. Operational and other real world constraints make the optimization process a rather complex problem. An efficient constraint handling method is especially important for this highly constrained problem. Artificial intelligence techniques such as Genetic Algorithms (GAs) have been introduced as a useful tool for the optimization of pavement maintenance problems. However, its efficiency largely depends on how constraints are handled. This paper discusses the Penalty Method as a constraint handling method for GA operations. A practical maintenance programming problem was considered to illustrate the method. How the penalty method works for the example problem is explained in details. The output of the GA operations is also analyzed from the pavement maintenance management point of view.
展开▼