Genetic programming: on the programming of computers by means of natural selection
Genetic programming: on the programming of computers by means of natural selection
Information theory
Genetic Algorithms in Search, Optimization and Machine Learning
Genetic Algorithms in Search, Optimization and Machine Learning
Numerical Optimization of Computer Models
Numerical Optimization of Computer Models
The Design of Innovation: Lessons from and for Competent Genetic Algorithms
The Design of Innovation: Lessons from and for Competent Genetic Algorithms
Schemata evolution and building blocks
Evolutionary Computation
Hybrid learning using genetic algorithms and decision trees for pattern classification
IJCAI'95 Proceedings of the 14th international joint conference on Artificial intelligence - Volume 1
Hi-index | 0.00 |
We use an information theory approach to investigate the role of mutation on Genetic Algorithms (GA). The concept of solution alleles representing information in the GA and the associated concept of information density, being the average frequency of solution alleles in the population, are introduced. Using these concepts, we show that mutation applied indiscriminately across the population has, on average, a detrimental effect on the accumulation of solution alleles within the population and hence the construction of the solution. Mutation is shown to reliably promote the accumulation of solution alleles only when it is targeted at individuals with a lower information density than the mutation source. When individuals with a lower information density than the mutation source are targeted for mutation, very high rates of mutation can be used. This significantly increases the diversity of alleles present in the population, while also increasing the average occurrence of solution alleles.