IBM ILOG Solver User's Manual > Evolutionary Algorithms > Using More Advanced EA Features > Using pool evaluators and partial replacement > Suggested Answers

Exercise 1

You should observe that increasing the selection pressure causes the optimal solution to be found more quickly. However, this can have a downside when more complex problems are considered. Suboptimal solutions can dominate any potentially promising (but currently inferior) solutions which emerge, resulting in a stagnation of the population. Genetic algorithms try to find a balance between diversification and intensification of the population by their choice of selection pressure.

Also note that even when the selection pressure is very low (use of a low power), solutions are still found quickly as you continue to perform partial replacement. This means that even if all solutions have a roughly equal chance of participating in a genetic operation, poorer solutions are more likely to be replaced by offspring. This effect has a tendency to increase solution quality.