Evolutionary Method of Detecting Intrusions using Different Population Dynamics
Sunitha Guruprasad1, Rio D’Souza G. L.2
1Sunitha Guruprasad*, Department of Computer Science and Engineering, St Joseph Engineering College, Mangaluru, India.
2Rio D’Souza G. L., Department of Computer Science and Engineering, St Joseph Engineering College, Mangaluru, India.
Manuscript received on February 10, 2020. | Revised Manuscript received on February 19, 2020. | Manuscript published on March 10, 2020. | PP: 1924-1931 | Volume-9 Issue-5, March 2020. | Retrieval Number: E2878039520/2020©BEIESP | DOI: 10.35940/ijitee.E2878.039520
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Chaos theory plays a vital role in any evolutionary based algorithms for avoiding the local optima and to improve the convergence speed. Various researchers have used different methods to increase the detection rate and to speed up the convergence. Some researchers have used evolutionary algorithms for the same purpose and has proved that the application of those algorithms provide good results. Most of the researchers have used population sizes which remains constant throughout the evolution. It has been seen that small population size may result in premature convergence and large population size requires more computation time to find a solution. In this paper, a novel application of different population dynamics to the genetic programming (GP) algorithm has been applied to manage the population size. The main focus was to improve the accuracy of the normal GP algorithm by varying the population sizes at each generation. The experiments were conducted on the standard GP algorithm using static and dynamic population sizes. Different population dynamics has been used to check the effectiveness of the proposed algorithm. The results obtained has shown that dynamic population size gives better results compared to static population size and also solves the problem of local optima.
Keywords: Detection Rate, Genetic Programming, Intrusion Detection, Population Dynamics, NSL-KDD, ISCX-2012, CICIDS2017.
Scope of the Article: Evolutionary Computing and Intelligent Systems