A genetic algorithm mimics evolution by natural selection. It begins with a set of solutions (i.e., a population) and allows some pairs of solutions, perhaps the best ones, to mate.
A crossover operation produces an offspring that inherits some characteristics of the parent solutions. At this point, the less desirable solutions are eliminated from the population so that only the fittest survive.
The process repeats for several generations, and the best solution in the resulting population is selected. Indicate how this algorithm can be viewed as examining a sequence of problem restrictions.
In what way does generation of offspring produce a relaxation of the current restriction?
If the "solution" of the relaxation is the selection of some good solutions, how does solution of the relaxation guide the selection of the next restriction?
Why is relaxation bounding, however, unhelpful in this algorithm?
Hint: relaxation bounding is helpful when it obviates the necessity of solving the current restriction. Think about how the current relaxation is obtained.