1、A brief introduction of artificial bee colony(ABC),E201102040杨丹,References,D. Karaboga, An idea based on honey bee swarm for numerical optimization. Technical Report-TR06, Erciyes University, Engineering Faculty, Computer Engineering Department, 2005 B. Basturk, D. Karaboga, An articial bee colony (
2、abc) algorithm for numeric function optimization, in: IEEE Swarm Intelligence Symposium 2006,Indianapolis, Indiana, USA, May 2006. D. Karaboga, B. Basturk, A powerful and efcient algorithm for numerical function optimization: articial bee colony (abc) algorithm,Journal of Global Optimization 39 (3)
3、(2007) 459471. D. Karaboga, B. Basturk, On the performance of articial bee colony (abc) algorithm, Applied Soft Computing 8 (1) (2008) 687697. D. Karaboga,B.Akay,A comparative study of artificial bee colony algorithm,Applied Mathematics and Computation 214 (2009) 108-132,1.INTRODUCTION,Two fundament
4、al concepts:self-organization and division of labour Self organization relies on four basic properties : positive feedback,negative feedback,fluctuations and multiple interactions Division of labour is believed to be more efficient than the sequential task performance. It also enables the swarm to r
5、espond to changed conditions in the search space.,2.BEHAVIOUR OF HONEY BEE SWARM,The emergence of collective intelligence of honey bee swarms consists of three essential components: Food sources Employed foragers Unemployed foragersthe model defines two leading modes of the behaviour: The recruiment
6、 to a nectar source The abandonment of a source,2.BEHAVIOUR OF HONEY BEE SWARM,There are two possible options for such a bee: i) It can be a scout and starts searching around the nest spontaneously for a food due to some internal motivation or possible external clue (S on Figure 1). ii) It can be a
7、recruit after watching the waggle dances and starts searching for a food source (R on Figure 1). After unloading the food, the bee has the following three options: i) It becomes an uncommitted follower after abandoning the food source (UF). ii) It dances and then recruits nest mates before returning
8、 to the same food source (EF1) iii) It continues to forage at the food source without recruiting other bees (EF2).,2.BEHAVIOUR OF HONEY BEE SWARM,Figure 1. The behaviour of honey bee foraging for nectar,2.BEHAVIOUR OF HONEY BEE SWARM,In the case of honey bees, the basic properties on which self orga
9、nization relies are as follows: i) Positive feedback: As the nectar amount of food sources increases, the number of onlookers visiting them increases, too. ii) Negative feedback: The exploitation process of poor food sources is stopped by bees. iii) Fluctuations: The scouts carry out a random search
10、 process for discovering new food sources. iv) Multiple interactions: Bees share their information about food sources with their nest mates on the dance area.,3.PROPOSED APPROACH,The main steps of the algorithm are given below: Send the scouts onto the initial food sources REPEAT Send the employed b
11、ees onto the food sources and determine their nectar amounts Calculate the probability value of the sources with which they are preferred by the onlooker bees Send the onlooker bees onto the food sources and determine their nectar amounts Stop the exploitation process of the sources exhausted by the
12、 bees Send the scouts into the search area for discovering new food sources, randomly memorize the best food source found so far UNTIL (requirements are met),3.PROPOSED APPROACH,An important control parameter of ABC:limitIf a solution representing a food source is not improved by a predetermined num
13、ber of trials, then that food source is abandoned by its employed bee and the employed bee is converted to a scout. The number of trials for releasing a food source is equal to the value of “limit “.,4.The performance of ABC algorithm,The probability with the food source located at will be chosen by
14、 a bee can be expressed as:,The position of the selected neighbour food source is calculated as the following:,4.The performance of ABC algorithm,Table 1 Numerical benchmark functions,4.The performance of ABC algorithm,Table 2 Mean of best function values obtained for 1000 cycle by ABC algorithm und
15、er different colony sizes,4.The performance of ABC algorithm,From Table 2 and it can be concluded that as the population size increases, the algorithm produces better results. However, after a sufficient value for colony size, any increment in the value does not improve the performance of the ABC al
16、gorithm significantly. For the test problems carried out in that work, the colony size of 50-100 can provide an acceptable convergence speed for search.,4.The performance of ABC algorithm,As mentioned before, the scout bee production is controlled by the control parameter limit in the ABC algorithm.
17、 There is an inverse proportionality between the value of limit and the scout production frequency. As the value of limit approaches to infinity, the total number of the scouts produced goes to zero. In order to show the effect of the scout production on the performance of the algorithm, the average
18、 of the production process best function values found for the different limit values (0.1 ne D,0.5 ne D, ne D and without scout) and colony sizes(20, 40 and 100) is given in Table 3. As seen from Table 3,for the multimodal functions f1,f3 and f4, when the scout production frequency is very high (lim
19、it value = 0.1 ne D) or zero (without scout), the results obtained by the ABC algorithm are worse than those produced by using the moderate values for limit, such as 0.5 ne D and ne D. For the unimodal functions f2 and f5,the production of scouts does not have any useful effect on the performance of
20、 the algorithm.,4.The performance of ABC algorithm,However, as expected, it improves the search ability of the algorithm for the multimodal functions and its benefit becomes much clearer for the smaller colony sizes.,4.The performance of ABC algorithm,Table 3 Effect of the limit value, which control
21、s the scout production, on the performance of the ABC algorithm (The bold value indicates the best among the values obtained under different limit values for the same function),5.A comparative study of ABC algorithm,In the field of evolutionary computation, it is common to compare different algorith
22、ms using a large test set, especially when the test involves function optimization. Attempting to design a perfect test set where all the functions are present in order to determine whether an algorithm is better than another for every function, is a fruitless task. That is the reason why, when an a
23、lgorithm is evaluated, we must look for the kind of problems where its performance is good, in order to characterize the type of problems for which the algorithm is suitable . We used 50 benchmark problems in order to test the performance of the GA, DE, PSO and the ABC algorithms. This set is large
24、enough to include many different kinds of problems such as unimodal, multimodal, regular, irregular,separable, non-separable and multidimensional.,5.A comparative study of ABC algorithm,In the last paper,the performance of ABC algorithm was compared with those of GA,PSO,DE and ES optimization algorithms. From the results obtained in this work, it can be concluded that the performance of ABC algorithm is better than or similar to that of these algorithms although it uses less control parameters and it can be efficently used for solving multimodal and multidimensional optimization problems.,
