Welcome to the InfoSci Platform
Reference Hub4
A Particle Swarm Optimization Approach for Reuse Guided Case Retrieval

A Particle Swarm Optimization Approach for Reuse Guided Case Retrieval

Nabila Nouaouria, Mounir Boukadoum, Robert Proulx
Copyright: © 2014 |Volume: 6 |Issue: 3 |Pages: 15
ISSN: 1942-9045|EISSN: 1942-9037|EISBN13: 9781466656826|DOI: 10.4018/ijssci.2014070102
Cite Article Cite Article

MLA

Nouaouria, Nabila, et al. "A Particle Swarm Optimization Approach for Reuse Guided Case Retrieval." IJSSCI vol.6, no.3 2014: pp.16-30. http://doi.org/10.4018/ijssci.2014070102

APA

Nouaouria, N., Boukadoum, M., & Proulx, R. (2014). A Particle Swarm Optimization Approach for Reuse Guided Case Retrieval. International Journal of Software Science and Computational Intelligence (IJSSCI), 6(3), 16-30. http://doi.org/10.4018/ijssci.2014070102

Chicago

Nouaouria, Nabila, Mounir Boukadoum, and Robert Proulx. "A Particle Swarm Optimization Approach for Reuse Guided Case Retrieval," International Journal of Software Science and Computational Intelligence (IJSSCI) 6, no.3: 16-30. http://doi.org/10.4018/ijssci.2014070102

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

The success of Case Based Reasoning (CBR) problem solving is mainly based on the recall process. The ideal CBR memory is one that simultaneously speeds up the retrieval step while improving the reuse of retrieved cases. In this paper, the authors present a novel associative memory model to perform the retrieval stage in a case based reasoning system. The described approach makes no prior assumption of a specific organization of the case memory, thus leading to a generic recall process. This is made possible by using Particle Swarm Optimization (PSO) to compute the neighborhood of a new problem, followed by direct access to the cases it contains. The fitness function of the PSO stage has a reuse semantic that combines similarity and adaptability as criteria for optimal case retrieval. The model was experimented on two proprietary databases and compared to the flat memory model for performance. The obtained results are very promising.

Request Access

You do not own this content. Please login to recommend this title to your institution's librarian or purchase it from the IGI Global bookstore.