The Cuckoo Search Algorithm is a recently developed nature inspired meta heuristic algorithm, which is established on the breeding behavior of Cuckoo species. Cuckoo search can be applied on a large variety of optimization problems. The main advantage of this search algorithm is its simplicity and better performance than many other agent or population based meta heuristic algorithms. The algorithm uses only one controlling parameter p, which makes it easier to implement and control. This parameter p, combined with the random walk mutations implemented by Lévy Flights, can control the performance and degree of exploration and exploitation of the algorithm. In this paper we have conducted a few experiments on Cuckoo Search algorithm with Lévy flights to discover the necessary conditions needed for the better performance of the algorithm. For this purpose we have taken different values of the controlling parameter p and observed the performance of the algorithm on benchmark problems, as well as its exploration and exploitation characteristics over different groups of benchmark functions.

The Flower Pollination Algorithm (FPA) a recently introduced bio-inspired algorithm based on the natural process of pollination of flowers. FPA has quickly drawn the attention of the research community due to its nature of both exploitation and exploration. In this paper we have introduced a novel variation in its mutation operation by linearly varying the value of its proximate probability parameter. We have evaluated the proposed algorithm on a number of benchmark problems and the experimental results are compared with the basic FPA. The empirical study indicates that the proposed FPA with variable exploration rate has better optimization performance than the basic FPA algorithm on most of the benchmark functions.