Evaluating Performance of Honey Bee Mating Optimization

Somayeh Karimi,Rahmat Sotudeh-Gharebagh,Navid Mostoufi

doi:10.1007/s10957-013-0336-2

Somayeh Karimi, Rahmat Sotudeh-Gharebagh + Show 1 more

Open Access

https://doi.org/10.1007/s10957-013-0336-2

Copy DOI

Abstract

The honey bee mating optimization (HBMO) algorithm is presented and tested with various test functions, and its performance is compared with the genetic algorithm (GA). It is shown that the HBMO algorithm can overcome the weaknesses of the GA. The HBMO converges faster than the GA. Even when the HMBO starts from a more improper initial condition than the GA, it can reach a better solution in a smaller number of function evaluations. Furthermore, in some cases, the GA was not able to reach the global minimum.

Highlights

A branch of nature-inspired algorithms, known as swarm intelligence, is focused on insect behavior in order to develop some meta-heuristics which can initiate the insect’s problem solution abilities
Honey bee mating algorithms (HBMO) belong to the novel swarm-based algorithms which are inspired by the marriage process in real bee colonies
For a required number of broods, a queen is selected in proportion to her fitness and is mated with a randomly selected sperm from her spermatheca. This process is similar to that of the genetic algorithm (GA), with the difference that in the GA each offspring is born from two parents while in the HBMO algorithm a brood may have some genes from one drone and some from another

Summary

Development of the HBMO Algorithm

There are five main steps in the development of the HBMO algorithm, as described below. For a required number of broods, a queen is selected in proportion to her fitness and is mated with a randomly selected sperm from her spermatheca This process is similar to that of the GA, with the difference that in the GA each offspring is born from two parents while in the HBMO algorithm a brood may have some genes from one drone and some from another. In algorithm The rate of improvement in the brood’s genotype, as a result of a heuristic application to that brood, is the definition of the fitness function for each worker. The remaining broods would not be killed but a predefined number of best broods (elites) would be selected and replaced with the worst ones In this way, the list of drones will be updated in each mating flight and with this replacement the exploitation will be powered. A new mating flight begins until all assigned mating flights are completed or convergence criteria are met

Genetic Algorithm

Result and Discussion

Conclusions