The growing need of finding efficient solutions to various optimization problems has motivated computer researchers to search for varied problem-solving methods. Nature has greatly inspired and motivated us in finding solutions to various optimization problems. Swarm Intelligence is a result of one such motivation where nature based swarm behavior of many social organisms like bees, bacteria, fireflies, cockroaches, mosquitoes is used for finding solutions to optimization problems. This paper introduces a new bio-Inspired optimization approach, namely, Pigeon Optimization Algorithm(POA) in the field of Swarm Intelligence. POA is based on the swarming behavior of passenger pigeons. The grouping and searching behavior of a flight of pigeons is used for finding the optimization solution to a given problem. The proposed algorithm demonstrates its suitability in finding the shortest path from a given source. The results obtained are compared and verified using Dijkstra's Algorithm. It has been found that the above algorithm has the potential and hence, can be used for solving different optimization problems in future.

New advancements in Meta-heuristics have forced the researchers to modify the existing algorithms in order to make them widely applicable to a large pool of complex problem set. Firefly Algorithm being a new nature-inspired algorithm has been used extensively for solving various optimization problems. The standard version namely, Standard Firefly Algorithm(SFA) was introduced in 2008 which uses the flashing behaviour of fireflies during night to obtain an optimization solution to a given problem. Two new modified variants of the SFA were introduced in 2012 which eliminated some of the limitations of the SFA. In this work, a new modified version of the current algorithm namely, New Modified Firefly Algorithm (NMFA) has been proposed and later its performance on various parameters is compared with SFA and its two more variants. Results demonstrate that the proposed algorithm is better in performance in comparison to all the other three algorithms when executed under a given set of control parameters. Other useful results of SFA and its variants are also tabulated.

No abstract

No abstract

No abstract