Clonal Algorithm for Dispatching Rule Selection

Clonal algorithm is based upon clonal evolutionary principles and selection theory. A population immune cell is explained with a fitness function for minimizing or maximizing the objective and as per to, each solution is evaluated. Subsequently, cloning of these random solutions is completed by copying them along with proliferation rate proportional to the fitness. After maturation, cloning of clones is completed through hyper-mutation process along with a rate inversely proportional to their fitness.

In short, the given steps are considered for applying CLONALG on any problem:

        I.            Randomly produce the initial population and calculate the population size and its maximum size.

       II.            For all immune cells, determine the value of fitness function.

      III.            For all immune cell of initial population, produce a number of clones based on its fitness or proliferation.

      IV.            Carry out the maturation process via randomly choosing two receptors or bits and swapping them.

      V.            Eliminate matured clones along with least fitness hence the new population contains the similar number of immune cells as there were initially.

      VI.            If new generation is equivalent to maximum generation, stop. Else go to iii.

On the basis of above steps, there are several following factors that are taken into consideration:

(a) Representation

(b) Initial population and alternative of fitness function

(c) Cloning

(d) Hyper-mutation operator

(e) Termination criteria

In such mechanism, a clonal algorithm is utilizing for enhancing the optimal dynamic dispatching strategy on the base of superior feature subset developing from the immune system or IS. Several crucial the steps of CLONALG that have been briefly discussed above are shown below.