other types of genetic algorithms

Steady State Genetic Algorithm:

A steady-state GA operates on overlapping populations in which only a subset of the current population is replaced in each generation.


Parallel Genetic Algorithm:

Apart from the implicit parallelism provided by schemata processing, the population-based approach of GAS makes them ideal for implementation on parallel, or networked, machines. Al though for many, the reason for using a parallet implementation will be a simple increase in speed due to the increase in computer power, others use pseudo-parallel approaches that are believed to improve performance even when used on sequential machines.

The basic idea behind most parallel programs is to divide a task into chunks and to solve the chunks simultaneously using multiple processors. Some parallelization methods use a single population, while other divides the population into several relatively isolated subpopulations. There are three main types of parallel GAs :
i) Global single-population Master slave GAs(Micro-grained model)
ii) Single-population fine grained (Cellular model)
iii) Multiple-Population coarse grained GAs (Island Model)

A. Master- Slave Scheme
In the master –slave scheme, as the name specifies one node becomes master and the other nodes become the slaves. The master node maintains the population of GA. The master node assigns individuals to slave nodes to parallelize the calculation of the fitness. The slave nodes calculate the fitness of each individual and sent it back to the master node and then the master nodes does the selection, crossover and mutation.

B. Cellular Scheme
In this scheme each individual is executed by a computing device separately. Individuals only mate with their connected neighbours during crossover and mutation. The neighbours of individuals are decided by the connection topology. The Cellular scheme is a fine grained parallel scheme.
C. Island Scheme
In the Island scheme, the population of GA is divided into several sub-population called islands. The sub-population works in isolation from each other. Island communicate and exchange some of their best individuals, this process is called migration. The migration helps to spread the local best solutions in islands to other islands, with this the population can evolve to better solutions