Bob McKay

Semantic similarity based crossover in GP: the case for real-valued function regression

In this paper we propose a new method for implementing the cross-over operator in Genetic Programming (GP) called Semantic Similarity based Crossover (SSC). This new operator is inspired by Semantic Aware Crossover (SAC) [20]. SSC extends SAC by adding semantics to control the change of the semantics of the individuals during the evolutionary process. The new crossover operator is then tested on a family of symbolic regression problems and compared with SAC as well as Standard Crossover (SC). The results from the experiments show that the change of the semantics (fitness) in the new SSC is smoother compared to SAC and SC. This leads to performance improvement in terms of percentage of successful runs and mean best fitness.

Intelligent and Evolutionary Systems

Artificial evolutionary systems are computer systems, inspired by ideas from natural evolution and related phenomena. The field has a long history, dating back to the earliest days of computer science, but it has only become an established scientific and engineering discipline since the 1990s, with packages for the commonest form, genetic algorithms, now widely available. Researchers in the Asia-Pacific region have participated strongly in the development of evolutionary systems, with a particular emphasis on the evolution of intelligent solutions to highly complex problems. The Asia-Pacific Symposia on Intelligent and Evolutionary Systems have been an important contributor to this growth in impact, since 1997 providing an annual forum for exchange and dissemination of ideas. Participants come primarily from East Asia and the Western Pacific, but contributions are welcomed from around the World. This volume features a selection of fourteen of the best papers from recent APSIES. They illustrate the breadth of research in the region, with applications ranging from business to medicine, from network optimization to the promotion of innovation.

Improving genetic classifiers with a boosting algorithm

We present a boosting genetic algorithm for classification rule discovery. The method is based on the iterative rule learning approach to genetic classifiers. The boosting mechanism increases the weight of those training instances that are not classified correctly by the new rules, so that in the next iteration the algorithm focuses the search on those rules that capture the misclassified or uncovered instances. We show that the boosted genetic classifier has higher accuracy for prediction, or from an alternative and perhaps more important perspective, uses less computational resources for similar accuracy, than the original genetic classifier.

The role of syntactic and semantic locality of crossover in genetic programming

This paper investigates the role of syntactic locality and semantic locality of crossover in Genetic Programming (GP). First we propose a novel crossover using syntactic locality, Syntactic Similarity based Crossover (SySC). We test this crossover on a number of real-valued symbolic regression problems. A comparison is undertaken with Standard Crossover (SC), and a recently proposed crossover for improving semantic locality, Semantic Similarity based Crossover (SSC). The metrics analysed include GP performance, GP code bloat and the effect on the ability of GP to generalise. The results show that improving syntactic locality reduces code bloat, and that leads to a slight improvement of the ability to generalise. By comparison, improving semantic locality significantly enhances GP performance, reduces code bloat and substantially improves the ability of GP to generalise. These results comfirm the more important role of semantic locality for crossover in GP.

A New Method for Simplifying Algebraic Expressions in Genetic Programming Called Equivalent Decision Simplification

Symbolic Regression is one of the most important applications of Genetic Programming, but these applications suffer from one of the key issues in Genetic Programming, namely bloat --- the uncontrolled growth of ineffective code segments, which do not contribute to the value of the function evolved, but complicate the evolutionary proces, and at minimum greatly increase the cost of evaluation. For a variety of reasons, reliable techniques to remove bloat are highly desirable --- to simplify the solutions generated at the end of runs, so that there is some chance of understanding them, to permit systematic study of the evolution of the effective core of the genotype, or even to perform simplification of expressions during the course of a run. This paper introduces an alternative approach, Equivalent Decision Simplification, in which subtrees are evaluated over the set of regression points; if the subtrees evaluate to the same values as known simple subtrees, they are replaced. The effectiveness of the proposed method is confirmed by computer simulation taking simple Symbolic Regression problems as examples.

A study on Genetic Programming with layered learning and incremental sampling

In this paper, we investigate the impact of a layered learning approach with incremental sampling on Genetic Programming (GP). The new system, called GPLL, is tested and compared with standard GP on twelve symbolic regression problems. While GPLL does not differ from standard GP on univariate target functions, it has better training efficiency on problems with bivariate targets. This indicates the potential usefulness of layered learning with incremental sampling in improving the efficiency of GP evolutionary learning.

Sampling bias in estimation of distribution algorithms for genetic programming using prototype trees

Probabilistic models are widely used in evolutionary and related algorithms. In Genetic Programming (GP), the Probabilistic Prototype Tree (PPT) is often used as a model representation. Drift due to sampling bias is a widely recognised problem, and may be serious, particularly in dependent probability models. While this has been closely studied in independent probability models, and more recently in probabilistic dependency models, it has received little attention in systems with strict dependence between probabilistic variables such as arise in PPT representation. Here, we investigate this issue, and present results suggesting that the drift effect in such models may be particularly severe - so severe as to cast doubt on their scalability.We present a preliminary analysis through a factor representation of the joint probability distribution. We suggest future directions for research aiming to overcome this problem.

Building on success in genetic programming: adaptive variation and developmental evaluation

We investigate a developmental tree-adjoining grammar guided genetic programming system (DTAG3P+), in which genetic operator application rates are adapted during evolution. We previously showed developmental evaluation could promote structured solutions and improve performance in symbolic regression problems. However testing on parity problems revealed an unanticipated problem, that good building blocks for early developmental stages might be lost in later stages of evolution. The adaptive variation rate in DTAG3P+ preserves good building blocks found in early search for later stages. It gives both good performance on small k-parity problems, and good scaling to large problems.

A self-adaptive mutations with multi-parent crossover evolutionary algorithm for solving function optimization problems

In this paper, we introduce a new self-adaptive evolutionary algorithm for solving function optimization problems. The capabilities of the new algorithm include: a) self-adaptive choice of Gaussian or Cauchy mutation to balance the local and global search on the variable subspace, b) using multi-parent crossover to exchange global search information, c) using the best individual to take place the worst individual selection strategy to reduce the selection pressure and ensure to find a global optimization. These enhancements increase the capabilities of the algorithm to solve Shekel problems in a more robust and universal way. This paper will present some results of numerical experiments which show that the new algorithm is more robust and universal than its competitors.

An Analysis of Semantic Aware Crossover

It is well-known that the crossover operator plays an important role in Genetic Programming (GP). In Standard Crossover (SC), semantics are not used to guide the selection of the crossover points, which are generated randomly. This lack of semantic information is the main cause of destructive effects from SC (e.g., children having lower fitness than their parents). Recently, we proposed a new semantic based crossover known GP called Semantic Aware Crossover (SAC) [25]. We show that SAC outperforms SC in solving a class of real-value symbolic regression problems. We clarify the effect of SAC on GP search in increasing the semantic diversity of the population, thus helping to reduce the destructive effects of crossover in GP.