Scheduling poses a difficult problem in numerous application areas. Realistic scheduling problems involve constraints that often cannot be precisely defined mathematically. As a result, traditional optimization methods are not always applicable for many scheduling and sequencing problems or are not as effective as we would like. Genetic algorithms not only offer a means of optimizing ill-structured problems, but also have the advantage of being a global search technique.
We have developed a genetic operator that generates high-quality solutions for sequencing or ordering problems; the performance of this operator is all the more remarkable because it does not require any heuristic or local optimization information. In fact, as is typical with genetic search, all that is required is that it be possible to obtain some overall evaluation of a sequence relative to other sequences. We refer to this operator as genetic edge recombination.
@incollection{ whitley91traveling,
author = "Darrell Whitley and Timothy Starkweather and Daniel Shaner",
title = "The Traveling Salesman and Sequence Scheduling: Quality Solutions Using Genetic Edge Recombination",
booktitle = "Handbook of Genetic Algorithms",
publisher = "Van Nostrand Reinhold",
address = "New York",
editor = "Lawrence Davis",
pages = "350--372",
year = "1991"
}
http://citeseer.ist.psu.edu/whitley90traveling.html
This paper gives a significantly more thorough treatment to Genetic Edge Recombination than RefWhitleyEtAl1989 does, but says essentially the same things.