Using the Genetic Algorithm to Find Snake-in-the-Box Codes

Genetic Algorithms are heuristic search schemes based on a model of Darwinian evolution. Although not guaranteed to find the optimal solution, genetic algorithms have been shown to be effective at finding near optimal and, in some cases, optimal solutions to combinatorially explosive problems. Finding a maximal length snake, a list of vertices satisfying specific constraints, in an n -dimensional hypercube, the "box", is the type of problem that suffers from (serious) combinatorial explosion. That is, as n increases, the size of the solution space that must be searched increases very rapidly. As a consequence, the maximum lengths of snakes have previously only been determined for n 6. Since their identification in the late 1950's, snakes (also known as Snake-In-The-Box codes) have received a significant amount of research attention. Snake-In-TheBox codes have a variety of applications in areas such as error-detection in analog-to-digital conversion, electronic combinat...

• Type: Conference
• Year: 1994
• Where: IEAAIE
• Authors: Walter D. Potter, Robert W. Robinson, John A. Miller, Krys Kochut, D. Z. Redys
• Keywords: Artificial Intelligence, Genetic Algorithm, Heuristic Search Schemes, IEAAIE 1994, Optimal Solutions