ISSN: 1813-0410

Keywords : ACS


A Proposal to Modulate ACS (Case study: Clustering Ionosphere database using FCM)

Soukaena Hassan Hashem

journal of kerbala university, Volume 11, Issue 1, Pages 205-211

This work trend to strength ACS, which is used as a solution for optimization problems, to make ACS more efficient and strongest to face the most important problems and faults in it is infrastructure, these problems are local optimum and stagnation. The enhancement is done by building proposed algorithm is called Modulated ACS, which is a trend to optimal solution by efficient treatment for local optimum and stagnation. PSO has been exploited as a solution to solve the both problems; that by modulates ACS by PSO parameters to converge the global and local updates. Several experiments are conducted to prove the robustness and strength of proposed Modulated ACS algorithm. The experimental works was done by applying traditional ACS and modulated ACS as an attribute reduction for Ionosphere database which include 351 instances, with Fuzzy C-Mean (FCM) clustering algorithm. The results show that; the modulated ACS introduces optimal time for finding the solutions than the traditional ACS. The modulated ACS reduces 50% of overall time spent by traditional ACS and the precision of clustering with traditional ACS was have 51 outlier instances between two clusters, where the precision of clustering with modulated ACS was have just 13 outlier instances between two clusters.

Advance ACS Using Chaos Searching Technique (Case Study: ACS-Based Network Routing Algorithms)

Soukaena Hassan Hashem

journal of kerbala university, Volume 10, Issue 0, Pages 67-74

In this paper a hybrid algorithm by combining the Ant Colony System (ACS) with Chaos Search (CS) is presented to enhance ACS (case study ACS-based network routing algorithms). The hybrid algorithm is injecting CS into ACS by initializing the algorithm with a set of random ants that travels in search space from source to destination. Then an optimization is obtained by CS to distinguish whither ant is feasible or not. That proposed model called (H-T-S-C). In each of iterations all feasible ants are ranked in ascending order. Ants in the front of the list are updated by ACS, while ants in the end of list are updated by CS. CS used here is not only to enhance the ants but also to improve the diversity of ant swarm so as to avoid ACS trapping the local optima. The results showed that the hybrid algorithm increased Message Delivery Ratio (MDR) about 10%, decreased jitter about 10%, decreased congestion about 5% and decreased time of search about 3%.