Geometric PSO + GP = Particle Swarm Programming

Genetic Programming, 2005

Particle Swarm Optimisers (PSOs) search using a set of interacting particles flying over the fitness landscape. These are typically controlled by forces that encourage each particle to fly back both towards the best point sampled by it and towards the swarm's best. Here we explore the possibility of evolving optimal force generating equations to control the particles in a PSO using genetic programming.

Journal of Artificial Evolution and Applications, 2008

Using a geometric framework for the interpretation of crossover of recent introduction, we show an intimate connection between particle swarm optimisation (PSO) and evolutionary algorithms. This connection enables us to generalise PSO to virtually any solution representation in a natural and straightforward way. The new geometric PSO (GPSO) applies naturally to both continuous and combinatorial spaces. We demonstrate this for the cases of Euclidean, Manhattan, and Hamming spaces and report extensive experimental results. We also demonstrate the applicability of GPSO to more challenging combinatorial spaces. The Sudoku puzzle is a perfect candidate to test new algorithmic ideas because it is entertaining and instructive as well as being a nontrivial constrained combinatorial problem. We apply GPSO to solve the Sudoku puzzle.

Genetic Programming, 2007

Using a geometric framework for the interpretation of crossover of recent introduction, we show an intimate connection between particle swarm optimization (PSO) and evolutionary algorithms. This connection enables us to generalize PSO to virtually any solution representation in a natural and straightforward way. We demonstrate this for the cases of Euclidean, Manhattan and Hamming spaces.

… of the 2005 conference on Genetic …, 2005

Particle Swarm Optimisation (PSO) uses a population of particles that fly over the fitness landscape in search of an optimal solution. The particles are controlled by forces that encourage each particle to fly back both towards the best point sampled by it and towards the swarm's best point, while its momentum tries to keep it moving in its current direction.

Using a geometric framework for the interpretation of crossover of recent introduction, we show an intimate connection between particle swarm optimisation (PSO) and evolutionary algorithms. This connection enables us to generalise PSO to virtually any solution representation in a natural and straightforward way. The new geometric PSO (GPSO) applies naturally to both continuous and combinatorial spaces. We demonstrate this for the cases of Euclidean, Manhattan, and Hamming spaces and report extensive experimental results. We also demonstrate the applicability of GPSO to more challenging combinatorial spaces. The Sudoku puzzle is a perfect candidate to test new algorithmic ideas because it is entertaining and instructive as well as being a nontrivial constrained combinatorial problem. We apply GPSO to solve the Sudoku puzzle.

Hybrid Artificial Intelligence Systems, 2010

Genetic Programming has emerged as an efficient algorithm for classification. It offers several prominent features like transparency, flexibility and efficient data modeling ability. However, GP requires long training times and suffers from increase in average population size during evolution. The aim of this paper is to introduce a framework to increase the accuracy of classifiers by performing a PSO based optimization approach. The proposed hybrid framework has been found efficient in increasing the accuracy of classifiers (expressed in the form of binary expression trees) in comparatively lesser number of function evaluations. The technique has been tested using five datasets from the UCI ML repository and found efficient.

Proceedings 2005 IEEE Swarm Intelligence Symposium, 2005. SIS 2005., 2005

Genetic programming (GP) is used to create fitness landscapes which highlight strengths and weaknesses of different types of PSO and to contrast population-based swarm approaches with non stochastic gradient followers (i.e. hill climbers). These automatically generated benchmark problems yield insights into the operation of PSOs, illustrate benefits and drawbacks of different population sizes and constriction (friction) coefficients, and reveal new swarm phenomena such as deception and the exploration/exploitation tradeoff. The method could be applied to any type of optimizer.

International Journal of Innovative Computing Information Control Ijicic, 2012

Genetic Programming (GP) is an emerging classification tool known for its flexibility, robustness and lucidity. However, GP suffers from a few limitations like long training time, bloat and lack of convergence. In this paper, we have proposed a hybrid technique that overcomes these drawbacks by improving the performance of GP evolved classifiers using Particle Swarm Optimization (PSO). This hybrid classification technique is a two-step process. In the first phase, we have used GP for evolution of arithmetic classifier expressions (ACE). In the second phase, we add weights to these expressions and optimize them using PSO. We have compared the performance of proposed framework (GPSO) with the GP classification technique over twelve benchmark data sets. The results conclude that the proposed optimization strategy outperforms GP with respect to classification accuracy and less computation.

ijicic.org

Genetic Programming (GP) is an emerging classification tool known for its flexibility, robustness and lucidity. However, GP suffers from a few limitations like long training time, bloat and lack of convergence. In this paper, we have proposed a hybrid technique that overcomes these drawbacks by improving the performance of GP evolved classifiers using Particle Swarm Optimization (PSO). This hybrid classification technique is a two-step process. In the first phase, we have used GP for evolution of arithmetic classifier expressions (ACE). In the second phase, we add weights to these expressions and optimize them using PSO. We have compared the performance of proposed framework (GPSO) with the GP classification technique over twelve benchmark data sets. The results conclude that the proposed optimization strategy outperforms GP with respect to classification accuracy and less computation.

This paper compares two evolutionary computation paradigms: genetic algorithms and particle swarm optimization. The operators of each paradigm are reviewed, focusing on how each affects search behavior in the problem space. The goals of the paper are to provide additional insights into how each paradigm works, and to suggest ways in which performance might be improved by incorporating features from one paradigm into the other.