PID controller optimized by Big Bang-Big Crunch algorithm for the evolution of router bandwidth demand
- PDF / 614,083 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 117 Downloads / 217 Views
ORIGINAL RESEARCH
PID controller optimized by Big Bang-Big Crunch algorithm for the evolution of router bandwidth demand Iordanis K. Giannopoulos1 • Asimakis K. Leros1 • Apostolos P. Leros2 Panagiotis Kofinas3 • Anastasios I. Dounis3
•
Received: 8 January 2019 / Accepted: 15 May 2019 Bharati Vidyapeeth’s Institute of Computer Applications and Management 2020
Abstract This paper presents an application of the Big Bang-Big Crunch algorithm to estimate with high precision the gains Kp, Ki and Kd of a PID controller. The optimized PID controller together with an evolution model is useful for the network administrators as it provides information within 5 or 15-min intervals for the bandwidth distribution. Furthermore, the Big Bang-Big Crunch algorithm results will be compared with those computed by the analysis of a classical PID controller. Finally, simulations with Simulink package are presented to compare the two methodologies. Keywords Big Bang-Big crunch algorithm PID tuning Network traffic modeling Router bandwidth demand
1 Introduction Routers are capable of providing periodic information about the network traffic such as cumulative value of packets in Gbps. Depending on the router settings, this information is released at the end of the intervals, usually, every 5 or even 15 min. This is done to prevent the router from occupying bandwidth and computational resources by sending this information within shorter intervals. A long interval reduces the workload of the router but creates a & Iordanis K. Giannopoulos [email protected] 1
Department of Information and Communication Systems Engineering, University of the Aegean, 83200 Samos, Greece
2
Department General, National and Kapodistrian University of Athens, Evia, Greece
3
Department of Industrial Design and Production Engineering, University of West Attica, Egaleo-Athens, Greece
blind spot on the on what is happening during this time. This is potentially problematic as 5 or 15 min is enough time for potential network threats to be implemented [1]. There is a need to have an estimated view of the traffic behavior during these intervals but without adding extra workload to the router. To achieve this, this paper proposes a novel approach by utilizing Proportional Integral Derivative (PID) [2] controller, tuned using big-bang bigcrunch (BB-BC) algorithm, explained in Sect. 4, to optimize the gains of the controller. The task described above is challenging because the traffic rate is depended on user behavior, hence it is nondeterministic. The signal can evolve in various ways and given that there is limited knowledge about the signal evolution it is difficult to predict its stochastic behavior, hence produce a reliable model. To overcome this problem, we use a PID controller, tuned by BB-BC algorithm. The PID controller can model the behavior of the signal evolution. However, this is not possible as the required Kp, Ki, Kd are unknown. BB-BC can optimize the controller gains and can produce values for the three optimal gains. In more detail,
Data Loading...
