The characterisation of genetic algorithm from a temporal point of view is quite interesting, and provides a description of the various objects in genetic algorithms offering novel insights into the dynamics of the process whereby the genetic algorithm reaches its objective: converging towards a global optimum, either a minimum or maximum depending on the problem at hand.
One of the problem faced is the need for a combination of a reasonable temporal logic and probabilistic logic in order to adequately characterise and describe both the temporal and stochastic aspects of the genetic algorithm. It is clear that such a formalism needs to be based on first order logic as well; an interval-based temporal logic seems to be best. For the probabilistic part we have taken the logic formulated by Bacchus ([106] ) as a basis.
Some work related to the present include e.g. the author's [512, 513, 514, 515] treating various aspects of time, selection and populations in genetic algorithm, as well as others: Leitch [368] with a fuzzy logic approach and Renders and Flasse [487] with hybrid methods for optimisation. Finally, we cite Fogel [187] who has dealt with genotype and phenotype mappings we use in the formalisation of the genetic algorithm from a temporal point of view.