Genetic programming: on the programming of computers by means of natural selection
Genetic programming: on the programming of computers by means of natural selection
A Survey of Optimization by Building and Using Probabilistic Models
Computational Optimization and Applications
Population-Based Incremental Learning: A Method for Integrating Genetic Search Based Function Optimization and Competitive Learning
Removing the Genetics from the Standard Genetic Algorithm
Removing the Genetics from the Standard Genetic Algorithm
Hi-index | 0.00 |
In this paper we discuss an application of simple stochastic optimization algorithm called the hill climbing with learning (HCwL) for a study of symbolic regression. A fundamental role in this approach plays the so-called probability vector w = (w1, w2, ..., wn) where an entry 0 ≤ w+i ≤1 specifies a probability that an i-th component of solution (e. g. a bit in binary representation) has a binary 1 value. An integral part of HCwL is a mutation process, where from a current solution xold is created a new solution xnew by a stochastic mutation process. The used probability vector w (considered here as a special type of collective memory) serves as an auxiliary device for a construction of new mutated solution xnew; in particular, it predicts promising directions during its creation that are specified by the previous history of adaptation process.