A New Index to Evaluate Solutions in the CLONALG Algorithm: Structural Similarity Index

Abstract

A new heuristic is proposed as a modification to the clonal selection based algorithm called the CLONALG of Artificial Immune Systems. In the CLONALG algorithm, the objective function values of the antibodies are used as an index for affinity, selection and mutation. It does not include a specified set of antigens to solve optimization problems. This mere use of objective function values hinders the utilization of the pattern recognition and adaptation ability of the Immune systems principles. Therefore, in the proposed heuristic, a Structural Similarity Index (SSI) is introduced as an additional criterion to facilitate the antibodies in exploiting the solution structures of antigens and clone accordingly. A hypothetical example problem from an automotive industry is solved to illustrate the application of the proposed heuristic and the results are compared with those obtained using the CLONALG algorithm. The comparison revealed that proposed solution methodology has a faster convergence pattern than the original CLONALG algorithm in solving the type of the underlying optimization problem.