Multi-chromosomes representations have been used in Genetic Algorithms to decompose complex solution representations into simpler components each of which is represented onto a single chromosome. This paper investigates the effects of distributing similar structures over a number of chromosomes. The solution representation of a simple mixed integer problem is encoded onto one, two, or three chromosomes to measure the effects. Initial results showed large differences, but further investigation showed that most of the differences were due to increased mutation, but multi-chromosome representation can give superior results.
Lecture Notes in Computer Science : Advanced Topics in Artificial Intelligence : Proceedings 10th Australian Joint Conference on Artificial Intelligence (AI 97), Perth, Western Australia, Australia, 30 November-04 December 1997,
Vol. 1342, p. 137-146