An effective hybrid cuckoo search algorithm with box-complex method for constrained optimization

It is well known that optimization is becoming an important tool for solving science and engineering problems. Mainly all the real world optimization problems are highly constrained in nature. Many times, solving such problems efficiently is a challenge task. Population based evolutionary optimization algorithms can potentially yield a global optimal solution. However, such evolutionary algorithms are designed for solving constrained optimization problems. Hence, effective constraint handling using

evolutionary techniques is still an active research area.



Cuckoo Search (CS) algorithm is a recently developed population based evolutionary optimization algorithm. While CS algorithm has been successfully applied for exploring the search space, it lacks good convergence property near the optimum point. On the other hand Box Complex (BC) method, a local search technique has good local convergence property. This paper presents a hybrid Cuckoo Search (CS) algorithm with effective constraint handling mechanism by Box Complex (BC) method. Each infeasible

member is processed using BC method by projecting it through the centroid of a fixed number of feasible members. This hybridization approach provides good convergence rate along with an effective constraint handling technique.



To demonstrate the efficacy of the proposed algorithm and BC method as effective constrained handling technique, a comparative study with the other popular algorithms and with popular conventional techniques under the cuckoo search framework has been presented using eleven test problems and four engineering design applications.