The performance of optimization algorithms is problem dependent and as per no free lunch theorem, there exists no such algorithm which can be efficiently applied to every type of problem(s). However, we can modify the algorithm/ technique in a manner such that it is able to deal with a maximum type of problems. In this study we have modified the structure of basic Artificial Bee Colony (ABC), a recently proposed metaheuristic algorithm based on the concept of swarm intelligence to optimize the models of software reliability. The modified variant of ABC is termed as balanced ABC (B-ABC). The simulated results show the efficiency and capability of the variant to solve such type of the problems.

In the basic Artificial Bee Colony (ABC) algorithm, if the fitness value associated with a food source is not improved for a certain number of specified trials then the corresponding bee becomes a scout to which a random value is assigned for finding the new food source. Basically, it is a mechanism of pulling out the candidate solution which may be entrapped in some local optimizer due to which its value is not improving. In the present study, we propose two new mechanisms for the movements of scout bees. In the first method, the scout bee follows a non-linear interpolated path while in the second one, scout bee follows Gaussian movement. Numerical results and statistical analysis of benchmark unconstrained, constrained and real life engineering design problems indicate that the proposed modifications enhance the performance of ABC.

Artificial Bee Colony (ABC), a recently proposed population based search heuristics which takes its inspiration from the intelligent foraging behavior of honey bees. In this study we have studied the impact of modification rate (MR) in basic ABC by gradually increasing it from 0.1 to 0.9. This impact is studied on four engineering design problems taken from literature. The simulated results show that it is beneficial to set the modification rate to a lower value.