Soft Computing Techniques for Optimal Capacitor Placement - Complex System Modelling and Control Through Intelligent Soft Computations - page 600

The na ï ve approach to solve any problem can be given as,

1. Formation of the search space: list all the feasible solutions of the given

problem,
2. Evaluate: evaluate their objective functions,
3. Best solution: choose the best solution amongst the various solutions.

flexible and capable of coping with more realistic
objective functions and constraints. They are used as a contrast to complete enu-
meration methods which guarantee to

These techniques are more fl

find the global optimum.
The various soft computing techniques used for optimum allocation of the
capacitors in the distribution system are discussed below.
Although, in principle, it is possible to solve any problem in this way, in practice
it is not due to the vast number of possible solutions to any real-world problem,
such as general optimal capacitor placement problem in distribution systems, of a
reasonable size.

Genetic Algorithms

Genetic Algorithms (GAs) are the soft computing tools, utilizing the concept of
evolution, to determine the optimal solution. It utilizes the survival of the fittest
concept to promote the growth of the healthier solutions in the search space, than
that of the unhealthier solutions.
Being a powerful tool for the optimization, several variants of the GA have been
proposed, viz., binary GA (BGA), real GA, niche GA, etc. These different forms of
GAs differ from one another in terms of how the solution space is processed, e.g., in
case of BGA the search space is decoded into binary form, whereas RGA processes
real form of data. In comparison to the traditional optimization methods, where the
solution moves from point to the other in the solution hyperspace, solutions in GA
search the solution hyperspace randomly (Aziz et al. 2013 ).
The various parameters that needs to be de

ned in a GA are,

1. Population size: It represents the random number of solutions in problem
hyperspace.
2. Crossover rate: It affects the rate at of crossover between the chromosomes. A
high probability represents introduction of new strings at a higher rate.
3. Mutation rate: It represents the probability of change in bit position of each
string in a new population after the selection process.

flowchart for the optimal placement and sizing of capacitor
banks in distribution system using GA (Masoum et al. 2004 ).
In spite of being a popular technique for optimization, GA lacks the

Figure 7 shows the fl

flexibility to
maintain a balance between global and local exploration of search space. Large
number of parameters needs to be decided to begin the solution to the problem, and
sometimes the convergence of the algorithm may fail (Boeringer and Werner 2004 ).

fl