Study on the Fuzzy Set and Ant Colony Algorithm Fused Algorithm Oriented at Fuzzy Data.doc

1、1Study on the Fuzzy Set and Ant Colony Algorithm Fused Algorithm Oriented at Fuzzy DataAbstract. In view of the increasingly large data size, the traditional algorithm model is no longer applicable. For this reason, a fuzzy set and ant colony optimization fused algorithm oriented at fuzzy data is in

2、troduced in this paper. Key words: Fuzzy Data; Fuzzy Set; Ant Colony Algorithm Introduction The fuzzy set and ant colony optimization fused algorithm is used as a new computing method. Different from the traditional algorithm, an accurate answer to question is not pursued in the algorithm, while unc

3、ertainty, inaccuracy, and partial authenticity universally exist in the fuzzy realities, so that problems can be easily processed, the cost will be lower, and large advantages can be displayed in general in the complex data processing. Brief introduction to soft computing Soft computing consists of

4、rough set, fuzzy logic, neural network, genetic algorithm, particle swarm optimization, and chaos. Compared with the traditional hard computing, soft 2computing is closer to the objective thing itself and the humans way of thinking, an also can more intelligently complete problem solving. The histor

5、ical development of soft computing is as shown in table 1. With the continuous development of related technologies, soft computing has developed very rapidly in recent years, and also has been widely applied to many fields. It plays an important role in complex system analysis, signal processing, pa

6、ttern recognition, fault diagnosis, data compression, etc. With the combination of soft computing with the technical knowledge of other fields, the specific and effective methods for solving problems are formed. In this paper, a fuzzy set and ant colony optimization fused algorithm is emphatically i

7、ntroduced. Introduction to fuzzy set theory and ant colony algorithm Fuzzy set theory In real life, not all things will inevitably occur under certain condition, and not all concepts are clear. There are a lot of fuzzy concepts in peoples thinking and language. For example, the vocabularies such as

8、smart, young, and beautiful have no clear boundaries, so these vocabularies become fuzzy concepts. The application of these fuzzy concepts is inaccurate, 3but can greatly simplify the amount of information, so that complicated problems become much easier to process. The theory solving the complex pr

9、oblems with the basic concept of this fuzzy set is called as fuzzy set theory, and generally can be divided into fuzzy mathematics, fuzzy logic and artificial intelligence, fuzzy system, uncertainty and information, and fuzzy decision, which are closely connected with each other. The historical deve

10、lopment of fuzzy set theory is as shown in table 2. Fuzzy set theory has been developed for more than 30 years, and has been widely applied. From the area of the social humanities to the area of engineering technology, fuzzy set theory has been widely used. Examples for the application of fuzzy set

11、theory are as shown in table 3. Ant colony algorithm Ant colony algorithm was proposed by Italian scholars in the 1990s. It is a new evolutionary algorithm. It was the earliest ant colony algorithm inspired from the ant system, while it was improved constantly by the researchers according to differe

12、nt understandings, so that the ant system owns many different versions of the ant colony algorithm. Moreover, these have also been used and optimized in various fields. The examples for the application of ant colony algorithm are as 4shown in table 4. In the ant colony algorithm, computation is impl

13、emented on the basis of distribution, and no a central control is available, and the communication occurs between distributed individuals, while this communication is indirect. Because of this characteristic, it is easier to combine with other optimization algorithms. The mutual coordination between

14、 simple individuals provides many new ideas for problem solving in the ant colony algorithm, and also the complex problems can help improve the ability of the algorithm in solution and response. These can make problem solving more optimized. Ant colony algorithm does not involve complex mathematical

15、 operations and has no higher requirements on the hardware and software of computer because it is relatively simple. These have greatly boosted the promotion of the algorithm. These are of very important theoretical and practical significance for the studies of ant colony algorithm. The real ant col

16、onies will release some information in the foraging path, while these pheromones will generate the shortest path between ant nest and food source so that nest and food source are connected better together. The ant colony algorithm is to simulate the real ant colonies. This characteristic is used as

17、a basis for 5implementing the algorithm. The ant is a type of social insects and intensively features sociality. Ants can complete a task better through collaboration between each other. This is because they are intensively collective. Thus, the disadvantage of a single ants too simple action is avo

18、ided, while the ant colony formed among simple individuals can give an expression to some extremely complex behaviors from the actions. Fuzzy set and ant colony algorithm fused algorithm If data objects are processed by an ordinary algorithm, probability transfer function is usually used as method,

19、but probability transfer function is a function using similarity as independent variables and thus will have some disadvantages. For example, the pick-up probability will be very low after the probability transfer function is substituted into the solving because of the great similarity if ants do no

20、t load anything and a data object is chosen, and also this data is very similar to the data within the neighborhood. However, the similar data objects will be moved if this probability is just greater than a randomly generated probability. Similarly, the lay-down probability of the data should be la

21、rger if ants load a data in somewhere, and also this data is very similar to the data in 6the neighborhood. Also, the similar objects will be laid down and moved if the lay-down probability is just less than a randomly generated probability. If ants are connected to the non-similar objects within th

22、e neighborhood during the unloading period, the objects should be picked up, but the non-similar objects will be picked up because the pick-up probability is just less than a randomly generated probability. It is assumed that the pheromones in all paths are equal in the initial moment, namely ij(0)

23、= C and ant k (k = 1, 2, 3,., m). In the process of movement, the shift direction of pheromones is decided by them according to the information amount of each path. Thus, at time t, the transfer probability of ant k from city i to city j is (t): Where, Jk(i)= 1，2，n- tabuk is a city allowed for ant k

24、 to choose in next step. In list tabuk, the cities k has gone by are recorded in the current iteration; k is not allowed to go by these cities in the circulation. If all n cities are added into tabuk, a cycle of travel is completed by k. Thus, the path of k can be a feasible solution to the problem

25、of TSP. In equation (1), the reciprocal between i and j is usually taken for ij; and respectively reflect the relative importance of accumulated information and inspiring information 7in the movement of ants. After a cycle of travel is completed by all ants, the pheromone in each path is changed as

26、follows: Where, (0 1) is the volatile coefficient of the pheromones in path; 1- is the persistent coefficient of pheromones; ij is the increment of the pheromones in the side (ij) of the current iteration. kij is the amount of the pheromones left by ant k on the side (ij) in the current iteration. If ant k does not go by side (ij), the value of kij is 0. kij is expressed as follows: