| 
|
Sexual Selection
Introduction
If you consider survival and reproduction as the main problems in the evolutionary of organisms and their behavior, simulating endogenous evolution should not only model natural selection (the difficulties of survival), but also sexual selection mechanisms (that model the difficulties in non-trivial reproduction).
If creatures have specific mating preferences (e.g. they like to mate with big partners), the creatures with this quality are expected to have a relatively greater number of offspring, and thus have greater fitness.
Sexual and natural selection can results in selection pressures in the same direction (both survival and reproduction benefit from being big), but also in pressures in opposite directions (mate preference towards bigger creatures, and chances of survival towards smaller creatures).
|
Sexual Selection
Modelling
One can easily model mate preference in Framsticks creatures, by making custom sensor neurons that (when the creature becomes fertile) select a partner-creature based on its properties (size, velocity, energy, relative location, etc.). This sensor gives activation that is proportional to the 'smell' of only this partner. The creature that has two of such sensors can move towards this partner in a Braitenberg style.
When this process leads to a collision between the two creatures (and they are still fertile), reproduction can be modelled depending on the specific experiment as mating with mutual agreement ('mating'), and without mutual agreement ('rape'). In the former case, an agreement can be modelled as a probability for the receiving creature to accept or decline a request for mating.
Creatures
The class of creatures that seems most relevant to this class of experiments are the 'Bugs'. This three-stick creatures have a Braitenberg style brain, and have two custom sensors that can be programmed to exhibit the mate preference mechanism described above.
However, to simulate the evolution of phonotaxis in cricket-like robots, for example, can best be modelled with the Khepera-like creature, since this is used in the original research.
Projects
There are many possible projects and experiments in this class of biological problems. Examples of problems that can be modelled and simulated include:
- 1. Morphology -- Can the interaction between sexual and natural selection lead to symmetric morphologies?
If sexual selection tends towards bigger creatures, and the task to Braitenberg-style navigation requires morphological symmetry, does this lead to bigger symmetrical creatures? And uinder what conditions?
- suggested reading:
- Maris, M. te. Boekhorst, R. (1996): Exploiting Physical Constraints: Heap Formation through Behavioral Error in a Group of Robots. IROS'96
- 2. Sympatric speciation through sexual selection -- Understanding speciation is a fundamental biological problem. It is believed that many species originated through allopatric divergence, where new species arise from geographically isolated populations of the same ancestral species. In contrast, the possibility of sympatric speciation (in which new species arise without geographical isolation) has often been dismissed, partly because of theoretical difficulties. Most previous models analysing sympatric speciation concentrated on particular aspects of the problem while neglecting others.
Can symaptric speciation (= spciation without geographical isolation) occur through assortative mating by evolution of mate preferences?
-
3. Parental Investment -- Trivers' (1972) theory of differential parental investment finally explained why males court females in most species. In biology, the differentiation of sexes is based on the relative size of their biological contribution to offspring. Those who contribute the greater portion are considered female, those who contribute the lesser portion are considered male.
-
4. Biodiversity through sexual selection -- What engenders biodiversity? Natural selection certainly adapts species to their ecological niches, but does it really create all of the new niches and new species to fill them? Consider: the most successful, complex, and numerous species on earth are composed of sexually-reproducing animals and flowering plants. Both groups typically undergo a form of sexual selection through mate choice: animals are selected by conspecifics and flowering plants are selected by heterospecific pollinators. This common feature suggests that the evolution of biodiversity may be driven not simply by natural-selective adaptation to ecological niches, but by subtle interactions between natural selection and sexual selection.
- 5. Sexual Selection and evolutionary search In standard GAs, individuals reproduce asexually: any two organisms can reproduce with each other. Gender separation and sexual selection inspire a model of gendered GAs in which crossover takes place between individuals from the opposite sex and the GA's evaluation, selection and mutation strategies depend on gender. Consequently, a pattern of cross-gender cooperation and intra-gender competition emerges. And a symbiotic relation between selection and crossover operations. Gendered GAs has been proven to significantly outperform standard GAs on various problems.
|
