Computational Intelligence

Chapter 12—Cas e Study Summaries 19

Ways to represent a scheduling problem span the spectrum from classic binary

coding to direct representation. We touch briefly on only a few methods. For more

information, see references such as Dimopoulos and Zalzala (2000).

The classic binary coding hearkens back to the canonical genetic algorithm

of Chapter 4. With parameters (machine numbers, task sequences, or whatever)

coded as 1s and 0s, this approach is seldom if ever used anymore, primarily because

the traditional crossover and mutation operators produce significant disruptions

to the feasibilit y of a schedule that must constantly be repaired to produce