The attempt to combine computer science and evolution begin in the end of the 1950s, but due to lack of a general coding scheme
People can only rely on variations and non-assignment to generate new gene structures, so they have little effect. By the mid-1960s, the United States
John Hol-Land at the University of Michigan in A. S. Fraser and H. J. Bremermann et al.
Bill string coding technology is proposed. This encoding is suitable for mutation operation, and is suitable for mating (ie hybrid) operations, and
Emphasis will be mating as the main genetic operation. Subsequently, Holland uses the algorithm for adaptiveness of natural and manual systems.
In the study of behavior, and published its pioneering books in 1975 "Adaptation in Natural
Andartificialsystems ". After that, Holland et al. promotes the algorithm, applied to optimization and machinery
In the problem of habits, and officially named genetic algorithm. General coding technology for genetic algorithms and simple and effective genetic operations
Laydate for its extensive, successful application.
Holland's genetic algorithm is often referred to as simple genetic algorithm (SGA), and its operation object is a group of two-string string (
Dyeing, individual), population. Here, each chromosome corresponds to a solution
. From the initial population, the selection policy based on adaptive value is used in the current population, and use hybridization.
And variations to produce the next generation population. Such a generation of evolution until the desired termination conditions are met. Need to point out
The current genetic algorithm is no longer limited to binary coding. Z. MichaleWicz will different encoding strategies (
That is, different data structures) The combination of genetic algorithms is called evolution program.
The core process of a typical genetic algorithm can be described below:
{
Random initialized population P (0) = {x1, x2, -, xn}, T: = 0;
Calculate the individuality of individuals in P (0);
While (do not meet the termination conditions) DO
{
Calculate the selection probability pi of the individual population according to the appropriateness of the individual and the selection strategy.
Select N1 (≤n) individual body according to Pi for genetic operation (copy, hybridization, and variation, etc.), after N1
Generate new population P (T 1) in replacing N1 in P (T)
Calculate the fitness of individuals in P (T 1), t = T 1;
}
}
The significance of some processes in the genetic algorithm and related algorithms (such as: replication process, miscellaneous
Equipment, etc.) and some specific nouns (such as variants, etc.) and the selection methods of each parameter [6
~ 8].
The genetic algorithm itself is a global random search method, which can design genetic algorithm in a certain way.
Ensure that it converges to the global optimal solution. This will overcome various numerical methods based on iterative principles on applicable issues.
Limits and cannot be guaranteed to converge to the overall optimal solution. For global optimization and analysis
