Viljem Žumer

Performance comparison of self-adaptive and adaptive differential evolution algorithms

Differential evolution (DE) has been shown to be a simple, yet powerful, evolutionary algorithm for global optimization for many real problems. Adaptation, especially self-adaptation, has been found to be highly beneficial for adjusting control parameters, especially when done without any user interaction. This paper presents differential evolution algorithms, which use different adaptive or self-adaptive mechanisms applied to the control parameters. Detailed performance comparisons of these algorithms on the benchmark functions are outlined.

Can a parser be generated from examples

One of the open problems in the area of domain-specific languages is how to make domain-specific language development easier for domain experts not versed in a programming language design. Possible approaches are to build a domain-specific language from parameterized building blocks or by language (grammar) induction. This paper uses an evolutionary approach to grammar induction. Grammar-specific genetic operators for crossover and mutation are proposed to achieve this task. Suitability of the approach is shown by small experiments where underlying grammars are successfully genetically obtained and parsers are than automatically generated.

An Analysis of the Control Parameters’ Adaptation in DE

The main goal of this chapter is to present an analysis of how self-adaptive control parameters are being changed during the current evolutionary process. We present a comparison of two distinct self-adaptive control parameters’ mechanisms, both using Differential Evolution (DE). The first mechanism has recently been proposed in the jDE algorithm, which uses self-adaptation for F and CR control parameters. In the second one, we integrated the well known self-adaptive mechanism from Evolution Strategies (ES) into the original DE algorithm, also for the F and CR control parameters. Both mechanisms keep the third DE control parameter NP fixed during the optimization process. They both use the same DE strategy, same mutation, crossover, and selection operations, even the same initial population, and they both use self-adaptation at individual level.

Implementation of multiple attribute grammar inheritance in the tool LISA

Multiple attribute grammar inheritance is a structural organization of attribute grammars where the attribute grammar inherits the specifications from ancestor attribute grammars, may add new specifications or may override some specifications from ancestor specifications. In the paper the implementation of multiple attribute grammar inheritance is described. The proposed approach is successfully implemented in the compiler/interpreter generator tool LISA ver. 2.0.

Extracting grammar from programs: brute force approach

Extracting grammar from programs attracts researchers from several fields such as software engineering, pattern recognition, computational linguistic and natural language acquisition. So far, only regular grammar induction has been successful, while purposeful context-free grammar induction is still elusive. We discuss the search space of context-free grammar induction and propose the Brute Force approach to the problem which, along with its various enhancements, can be further used in collaboration with the Evolutionary approach, as described in the accompanying paper.

LISA: a tool for automatic language implementation

In the paper a tool for automatic language implementation is described. From formal language definition LISA1 produces efficient interpreters or compilers in C++. For lexical and syntax analysis well known formal methods: regular expressions and BNF are used. Semantics of the defined language is described with attribute grammars. LISA is a part of the ongoing effort to build an environment that supports experimentation in automatic language implementation for educational and industrial purposes.

Automatic Generation of Language-based Tools

Many tools can be automatically derived from formal language definitions, such as compilers/interpreters, editors, analyzers, visualizers/animators, etc. Some examples of language-based tools generated automatically by the LISA system are described in the paper. In addition the specification of an algorithm animator and program visualizer, Alma, generated from an extended LISA input-grammar is discussed; LISA principles and code are reused in Alma implementation.

History mechanism supported differential evolution for chess evaluation function tuning

This paper presents a differential evolution (DE) based approach to chess evaluation function tuning. DE with opposition-based optimization is employed and upgraded with a history mechanism to improve the evaluation of individuals and the tuning process. The general idea is based on individual evaluations according to played games through several generations and different environments. We introduce a new history mechanism which uses an auxiliary population containing good individuals. This new mechanism ensures that good individuals remain within the evolutionary process, even though they died several generations back and later can be brought back into the evolutionary process. In such a manner the evaluation of individuals is improved and consequently the whole tuning process.

An Adaptive Differential Evolution Algorithm with Opposition-Based Mechanisms, Applied to the Tuning of a Chess Program

This chapter describes an algorithm for the tuning of a chess program which is based on Differential Evolution using adaptation and opposition based optimization mechanisms. The mutation control parameter F is adapted according to the deviation of search parameters in each generation. Opposition-based optimization is included in the initialization, and in the evolutionary process itself. In order to demonstrate the behaviour of our algorithm we tuned our BBChess chess program with a combination of adaptive and opposition-based optimization. Tuning results show that adaptive optimization with an opposition-based mechanism increases the robustness of the algorithm and has a comparable convergence to the algorithm which uses only adaptation optimization.

Software complexity—an alternative view

A different view on software complexity based on long range correlations is presented in this paper. Four different methods are proposed first and then used in the analysis of twenty FORTRAN programs. All methods being used indicate the presence of long range correlations, but every method gives different results. The presence of long range correlations denotes that we can infact regard the complexity of computer programs from a different perspective.