Question

List advantages, disadvantages, in what areas/applications used efficiently for the following:

Evolutionary algorithms.

Evolutionary strategies.

Evolutionary programming.

Genetic, simulated annealing, tabu search, ant colony algorithms

Answer 1

Evolutionary algorithm - The concept of evolutionary algorithm, as the name suggests, uses the biological principle of evolution to find the optimal solution to a difficult problem. The three main priciples used to arrive at the best solution from the set of all possible inputs (Solution Space), are as follows:

Reproduction - is inspired by the role of reproduction in natural evolution.Existing solutions are combined to come up with new and improved solutions. The new solutions have some features of each parent.

Mutation - is inspired by the role of mutation of an organism's DNA in natural evolution. The evolution algorithm periodically tries to make random changes in one or more solutions of the current population, in order to come up with new solution.

Survival of the fittest - is inspired by the role of natural selection in evolution. Here, an evolutionary algorithm selects the most fit (most feasible) solutions from the population and eliminates the rest of the solutions.

Advantages:

1. EA only need little problem specific knowledge and that they can be applied on a broad range of problems.
2. EA can be applied on complex problems with discontinuous, non-differentiable and possibly noisy target functions.
3. EA is best suited for problems where multiple solutions are required (Multi-objective optimization problems).
4. Evolutionary Algorithms support parallel implementation.

Disadvantages:

Due to their random nature, evolutionary algorithms are never guaranteed to find an optimal solution for any problem.

Using population approach can be expensive. (can be overcome with parallel implementation)

Evolutionary algorithm can keep running for infinite amount of time or iterations and doesn't know when to stop.

Areas/Applications: discrete or integer optimization problems

Evolutionary Strategies - also simulates the three principles of evolution like Evolutionary Algorithm. However, unlike Evolutionary Algorithm where a solution is represented using a sequence of integers, Evolution Strategy uses a sequence of real numbers.

Advantages:

Evolutionary Strategiess are best suited for continuous parameter optimization problems associated with laboratory experiments.

Evolutionary Programming: Like both EA and ES, EP is a useful method of OPTIMIZATION when other techniques are not possible. However, instead of simulating the principles of evolution, it emphasizes on the behavioral linkage between parents and children.