What does Evolutionary Algorithm mean?
An evolutionary algorithm is considered a component of evolutionary computation in artificial intelligence. An evolutionary algorithm functions through the selection process in which the least fit members of the population set are eliminated, whereas the fit members are allowed to survive and continue until better solutions are determined. In other words, evolutionary algorithms are computer applications which mimic biological processes in order to solve complex problems. Over time, the successful members evolve to present the optimized solution to the problem.
Techopedia explains Evolutionary Algorithm
Evolutionary algorithms make use of concepts in biology such as selection, reproduction and mutation. There are three basic types of evolutionary algorithms, namely:
- Genetic algorithms
- Evolutionary programming
- Evolutionary strategies
Unlike traditional optimization techniques, evolutionary algorithms depend on random sampling. An evolutionary algorithm has a population of candidate solutions, unlike classical methods, which try to maintain a single best solution. There are two prerequisites associated with evolutionary algorithms:
- The candidate solutions need to be encoded to the issue.
- A fitness function needs to return a score between 1 and 100 so that evolutionary algorithms can be better to be applied to the problems.
There are many benefits associated with evolutionary algorithms. One of the biggest advantages comes in the flexibility gains, as most evolutionary algorithm concepts are adaptable to even complex problems. Most evolutionary algorithms are fit to meet the objective target as well. Better optimization is possible with evolutionary algorithms, as the population of solutions prevents the algorithm from getting locked in a particular solution.
There are a few drawbacks associated with evolutionary algorithms. For one, the solution provided by an evolutionary algorithm is only better while compared to other known solutions. As such, the algorithm cannot prove that any solution is totally optimal, only that it is optimal compared to the other results.