Telecommunications network design algorithms
Telecommunications network design algorithms
An approach to a problem in network design using genetic algorithms
An approach to a problem in network design using genetic algorithms
A weighted coding in a genetic algorithm for the degree-constrained minimum spanning tree problem
SAC '00 Proceedings of the 2000 ACM symposium on Applied computing - Volume 1
Computers and Intractability: A Guide to the Theory of NP-Completeness
Computers and Intractability: A Guide to the Theory of NP-Completeness
A Genetic Algorithm for Survivable Network Design
Proceedings of the 5th International Conference on Genetic Algorithms
PPSN VI Proceedings of the 6th International Conference on Parallel Problem Solving from Nature
A New Genetic Algorithm for the Optimal Communication Spanning Tree Problem
AE '99 Selected Papers from the 4th European Conference on Artificial Evolution
Proceedings of the 6th International Conference on Genetic Algorithms
The gambler's ruin problem, genetic algorithms, and the sizing of populations
Evolutionary Computation
Redundant representations in evolutionary computation
Evolutionary Computation
On Optimal Solutions for the Optimal Communication Spanning Tree Problem
Operations Research
The property analysis of evolutionary algorithms applied to spanning tree problems
Applied Intelligence
On a property analysis of representations for spanning tree problems
EA'05 Proceedings of the 7th international conference on Artificial Evolution
| Hi-index | 0.00 |
When using genetic and evolutionary algorithms (GEAs) for the optimal communication spanning tree problem, the design of a suitable tree network encoding is crucial for finding good solutions. The link and node biased (LNB) encoding represents the structure of a tree network using a weighted vector and allows the GEA to distinguish between the importance of the nodes and links in the network. This paper investigates whether the encoding is unbiased in the sense that all trees are equally represented, and how the parameters of the encoding influence the bias. If the optimal solution is underrepresented in the population, a reduction in the GEA performance is unavoidable. The investigation reveals that the commonly used simpler version of the encoding is biased towards star networks, and that the initial population is dominated by only a few individuals. The more costly link-and-node-biased encoding uses not only a node-specific bias, but also a link-specific bias. Similarly to the node-biased encoding, the link-and-node-biased encoding is also biased towards star networks, especially when using a low weighting for the link-specific bias. The results show that by increasing the link-specific bias, that the overall bias of the encoding is reduced. If researchers want to use the LNB encoding, and they are interested in having an unbiased representation, they should use higher values for the weight of the link-specific bias. Nevertheless, they should also be aware of the limitations of the LNB encoding when using it for encoding tree problems. The encoding could be a good choice for the optimal communication spanning tree problem as the optimal solutions tend to be more star-like. However, for general tree problems the encoding should be used carefully.