27-07-2012, 11:05 AM
Genetic Algorithm Optimization of Wireless Communication Networks
Genetic Algorithm Optimization.pdf (Size: 275.25 KB / Downloads: 42)
Introduction
The proliferation of wireless communication networks,
particularly low cost configurations, has brought with it a natural interest in
network optimization [I]. Specifically, the low cost emphasis in many modem
wireless networks often places severe constraints on the acceptable level of
transceiver processing circuitry complexity. Available transmitter power is also
often at a premium in these modem networks. The solution to the design problem
therefore usually requires significant optimization coupled with network
simulation. This paper considers the synthesis of a 2-D network of N transceiver
nodes with a known spatial distribution using a Genetic Algorithm optimization
method. Each of the nodes consists of a relatively simple a transceiver (antennas, a
receiver and a transmitter) in which only limited adjustments of antenna pattem and
transmission power level are available. The goal of the optimization is to maximize
nodal signal to noise ratios (SNR) while minimizing the transmitter power levels.
The Genetic Algorithm Optimizer
As noted above, the optimization strategy applied to the network design probleh in this paper is the Genetic Algorithm. Genetic Algorithm (GA) optimizers are robust, stochastic search methods modeled on natural selection and evolution found in nature. As an
optimizer, the powerful heuristic of GA is effective at solving complex,
combinatorial problems. GA optimizers are particularly effective when the goal is to find an approximate global maxima in a high dimension, multi-modal function domain in a near optimal manner. GAS differ from more traditional techniques in that they operate on a group (or popdation) of trial solutions in parallel, they normally operate on a coding of the function parameters (chromosome) rather than
the parameters themselves and they use simple, stochastic operators (selection, crossover, and mutation) to explore the solution domain. In keeping with the natural world analogy, successive population of trial solutions are called generation. Subsequent generations are made up of children produced through the selective reproduction of pairs ofparents taken from the current generation.
Optimization Fitness Function and Codings
The object function, or fitness function, used to assign a fitness value to each of the individuals in the GA
population is the only link between the physical problem being optimized and the
general GA machinery. The only constraint on the form and content of the object
function imposed by the GA is that the fitness value returned by the object function
is in some manner proportional to the "desirability" of a given trial solution
represented as a set of input parameters to the object function. The chromosome is
a string encoding a potential solution vector and represents the individual. In the
case of network topology optimization, the fitness function consists of a calculation
of the average path length for a given ordering of the nodes.