Use
Benchmark data
The tables below are taken verbatim from the TR06 (2005) technical report in which the ABC algorithm was first introduced. Each experiment was repeated 30 times with different random seeds and the average function values of the best solutions were recorded.
Benchmark functions and results
| Function | Range | Global minimum | Mean | Std. dev. |
|---|---|---|---|---|
| Sphere (5 dimensions)Continuous, convex and unimodal function. | [-100, 100] | f(0) = 0 | 4.45E-17 | 1.13E-17 |
| Rosenbrock (2 dimensions)The global optimum lies inside a long, narrow, parabolic valley; the variables are strongly dependent. | [-2.048, 2.048] | f(1,1) = 0 | 0.002234 | 0.002645 |
| Rastrigin (10 dimensions)Cosine modulation added to the Sphere function produces many local minima. | [-600, 600] | f(0) = 0 | 4.68E-17 | 2.64E-17 |
Control parameters
| Parameter | Description | Value |
|---|---|---|
| swarmsize | Colony size | 20 |
| limit | Failed trials before a source is abandoned | number of onlooker bees × dimension |
| onlooker bees | Share of the colony | 50% |
| employed bees | Share of the colony | 50% |
| scout bees | Per cycle | 1 |
| maximum cycles | Stopping criterion (MCN) | 2000 |
Source: TR06 — An Idea Based on Honey Bee Swarm for Numerical Optimization (2005)