Encoding, Initialization and Cloning

A population of antibodies can be initialized employing binary, real or integer strings depending upon the problem. The selection of encoding strategy depends upon the difficult complexity and performance bound. For minute size problems, binary representation produces great results. For average size problems, integer coding is rather often employed and for complex problems are extremely large, real coding can be employed. As far as performance bound is related, clonal algorithm is expected to offer exploration of in an raising order, from binary coding to integer coding to real coding. Conversely, it is not all the time true owing to stochastic nature of the algorithm. Various types of encoding strategies are shown in following Figure.

Rather often, lymphocytes are referred to like the antibody for the computational causes and no distinction is made among a B-cell and its antibody.

In case of constrained optimization, there is all the time a possibility of constraint violation during string initialization and hyper-mutation. Thus, each string is checked for constraint violation and repaired hence all the members of initial population are feasible. Affinity is computed via fitness values such are computed employing equation can be accompanied by a penalty function as represented in equation 20. In addition to repairing strategy, there are refusing and penalty strategy as well that are discussed later. All of the antibodies from the initial pool are copied into a fixed number of clones to produce a temporary population of clones. Such type of population of clones is made to undergo maturation process via hyper-mutation mechanism followed through their affinity evaluation and repair in case of any constraint violation.

Following figure shows a schematic diagram implying the steps in resolving a parameter optimization problem in computer integrated manufacturing. It is significant to note that a new population of the similar size as primary population of the antibodies is chosen from the mutated clones and this completes the initial iteration. In the next iteration, such fresh population is made to undergo cloning and hyper-mutation as conversed above and likewise.

                                                                            Figure: Flow Chart for Solving a parameter Optimization Problem in CIM