Professional Documents
Culture Documents
Abstract— In this paper, we introduce a novel controllers such as Genetic Algorithms and ant colony
active queue management scheme based on a Proportional- optimization.
Integral-Derivative controller with ant colony optimization
for the Internet system. The ACO algorithm searching The paper is organized as follows: the second section is
technique is applied to search for the best gain parameters of dedicated to the presentation of the TCP/AQM problem
the PID controller among a stability area. This stability addressed in this study. In the third section we explain the
region is determined using an extension of Hermite-Biehler PID tuning approach. The fourth section is about the Ant
theorem applied to delayed systems. The step responses of Colony Optimization (ACO) technique. Section five
closed loop system of the PID controller with the ACO tackles the Multi-objective Optimization concept and the
algorithm are compared with Genetic Algorithm (GA) and design of a multi-objective ant colony optimization
Ziegler-Nichols (ZN) methods. Simulation results show the implemented to the tuned PID controller. Finally, we
performance and effectiveness of using ACO algorithm for provide some simulations to show the performance of the
PID tuning. proposed method.
Keywords —TCP; AQM; MultiObjective Optimization; Ant II. PROBLEM STATEMENT
Colony Optimization; PID; Delay Systems;
Internet is comprised by the interconnection of highly
I. INTRODUCTION heterogeneous, complex and dynamic modules and the
dominant transport protocol for the Internet is TCP, It
The public internet is a computer network, that is, a adopts the end-to-end windows-based flow control to avoid
network that interconnects computers around the world congestion.
using a common software protocol for transmitting and
receiving data. This protocol is known as TCP/IP. TCP’s The main property of TCP is to guarantee reliable
congestion control algorithm has been described as the communication on the Internet. Congestion occurs when a
largest human made feedback controlled system in the link or node is carrying so much data that its quality of
world [1]. service deteriorates. To correct the problems with tail-drop
buffer management, Floyd and Jacobsen introduced the
Congestion is one of the most important problems concept of AQM, in which the routing nodes use a more
encountered in communication networks. It occurs when a sophisticated algorithm called RED to manage their queues
link or node carries so much data that its quality of service [4]. Different algorithms was proposed and introduced into
deteriorates. This results in queue delay, packet loss or the Internet routers such as BLUE, REM or AVQ [2], [5].
blocking of new connections, resulting in low throughput
and possibly congestion collapse. Most research in this area We design a congestion control mechanism based on
try to avoid these situations of congestion. The introduction Active Queue Management (AQM) techniques by using
of the concept of AQM provides a solution to the problem control theory. In order to use control theory, we propose
in which routers actively drop packets from queues which to introduce a mathematic model of the TCP behavior
in used to keep the packet queue size correct. Recently, which has been developed by Misra et al, and which has
various AQM schemes have been proposed in the literature been simplified by Hollot et al [6], [7].
to increase network utilization by regulating queues at the This dynamic model of TCP/AQM is based on fluid-
bottleneck links in TCP/IP networks [2], [3]. flow and stochastic differential equations while ignoring
The automatic control technique that has received more the TCP timeout and the slow-start mechanism. The non-
attention is the PID controller. It is widely used in various linear model used for the study is given by the following
engineering applications. Despite its simplicity, it can two coupled non-linear differential equations:
successfully solve a variety of complex problems. But it is
often difficult to find optimal PID parameters with 1 ( ). − ( )
( )= − − ( ) (1)
traditional techniques. Several new methods have been ( ) 2 − ( )
developed recently to tune the parameters of the PID ( )
( )= ( ) − (2)
( )
where w(t) denotes the average of TCP windows size step is essential for any design and tuning of PID
(packets), q(t) the average of queue length controllers. In the course of numerous research studies
( ) dedicated to the PID regulator, several configurations have
(packets), ( ) = + the round-trip time (sec),
been assigned to it. The most common of these is the closed
the propagation delay, C the router’s transmission capacity, loop SISO [14] which is shown in Figure 1.
N(t) the number of TCP sessions. ( ) and ( ) denote
respectively the time-derivative of ( ) and ( ). ( ) ( )
C(s): PID
u( ) ( )
+ G(s): Plant
Most systems in the real world are nonlinear, then the - Controller
analysis of certain required properties and design
controllers become difficult, if not impossible. For this
reason the linearization of those models using some
approximation become a relevant step. Taking (w, q) as the Fig.1. Feedback control system.
state and p as input, the equations linearization around an
where , , and are respectively the desired, error,
operating point gives a two-order model with delay that is
command and output signals, C(s) is the PID controller and
given by the following transfer function [8-10]:
G(s) is the plant to be controlled.
( )= (3) The control law of the PID regulator can be written in
2
+ + +2+ Laplace transform as:
Generally, industrial process are modelled as first order ( )
( )= = + + (5)
system, which is comparatively easy to analysis: ( )
G( ) = (4) where , and are respectively the proportional,
+1 integral and derivative gains of the PID controller.
where
IV. ANT COLONY OPTIMIZATION
= is the state gain, The Ant Colony Optimization (ACO) came into being
as a result of the ant behavior in nature when searching for
= + − + 2 is the constant time, food. These social insects use, to communicate, a volatile
substance called pheromone, which they use to mark their
and route between the nest and the food.
= +2 − is the time delay of the plant.
ISECONCLUSION
0.2382 0.0551 0.1051
GA-PID
ITAE 0.2210 0.0415 0.0618