鸡群算法

Halcom

鸡群算法。。。。。

1. % -------------------------------------------------------------------------
   
2. % Chicken Swarm Optimization (CSO) (demo)
   
3. % Programmed by Xian-Bing Meng
   
4. % Updated at Jun 21, 2015.   
   
5. % Email: x.b.meng12@gmail.com
   
6. %
   
7. % This is a simple demo version only implemented the basic idea of CSO for        
   
8. % solving the unconstrained problem, namely Sphere function.   
   
9. % The details about CSO are illustratred in the following paper.                                                
   
10. % Xian-Bing Meng, et al. A new bio-inspired algorithm: Chicken Swarm
    
11. %    Optimization. The Fifth International Conference on Swarm Intelligence
    
12. %
    
13. % The parameters in CSO are presented as follows.
    
14. % FitFunc    % The objective function
    
15. % M          % Maxmimal generations (iterations)
    
16. % pop        % Population size
    
17. % dim        % Dimension
    
18. % G          % How often the chicken swarm can be updated.
    
19. % rPercent   % The population size of roosters accounts for "rPercent"
    
20. %   percent of the total population size
    
21. % hPercent   % The population size of hens accounts for "hPercent" percent
    
22. %  of the total population size
    
23. % mPercent   % The population size of mother hens accounts for "mPercent"
    
24. %  percent of the population size of hens
    
25. %
    
26. % Using the default value,CSO can be executed using the following code.
    
27. % [ bestX, fMin ] = CSO
    
28. % -------------------------------------------------------------------------
    
29. 
    
30. %*************************************************************************
    
31. % Revision 1
    
32. % Revised at May 23, 2015
    
33. % 1.Note that the previous version of CSO doen't consider the situation
    
34. %   that there maybe not exist hens in a group.
    
35. %   We assume there exist at least one hen in each group.
    
36. 
    
37. % Revision 2
    
38. % Revised at Jun 24, 2015
    
39. % 1.Correct an error at line "100".
    
40. %*************************************************************************
    
41. 
    
42. % Main programs
    
43. 
    
44. function [ bestX, fMin ] = CSO( FitFunc, M, pop, dim, G, rPercent, hPercent, mPercent )
    
45. % Display help
    
46. help CSO.m
    
47. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
    
48. % set the default parameters
    
49. if nargin < 1
    
50.     FitFunc = @Sphere;
    
51.     M = 1000;  
    
52.     pop = 100;  
    
53.     dim = 20;  
    
54.     G = 10;      
    
55.     rPercent = 0.15;
    
56.     hPercent = 0.7;  
    
57.     mPercent = 0.5;                  
    
58. end
    
59. 
    
60. rNum = round( pop * rPercent );    % The population size of roosters
    
61. hNum = round( pop * hPercent );    % The population size of hens
    
62. cNum = pop - rNum - hNum;          % The population size of chicks
    
63. mNum = round( hNum * mPercent );   % The population size of mother hens
    
64. 
    
65. lb= -100*ones( 1,dim );   % Lower bounds
    
66. ub= 100*ones( 1,dim );    % Upper bounds
    
67. 
    
68. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
    
69. %Initialization
    
70. for i = 1 : pop
    
71.     x( i, : ) = lb + (ub - lb) .* rand( 1, dim );
    
72.     fit( i ) = FitFunc( x( i, : ) );
    
73. end
    
74. pFit = fit; % The individual's best fitness value
    
75. pX = x;     % The individual's best position corresponding to the pFit
    
76. 
    
77. [ fMin, bestIndex ] = min( fit );        % fMin denotes the global optimum
    
78. % bestX denotes the position corresponding to fMin
    
79. bestX = x( bestIndex, : );   
    
80. 
    
81. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
    
82. % Start the iteration.
    
83. 
    
84. for t = 1 : M
    
85.    
    
86.     % This parameter is to describe how closely the chicks would follow
    
87.     % their mother to forage for food. In fact, there exist cNum chicks,
    
88.     % thus only cNum values of FL would be used.
    
89.     FL = rand( pop, 1 ) .* 0.4 + 0.5;  
    
90.    
    
91.     % The chicken swarm'status about hierarchal order, dominance  
    
92.     % relationship, mother-child relationship, the roosters, hens and the
    
93.     % chicks in a group will remain stable during a period of time. These  
    
94.     % statuses can be updated every several (G) time steps.The parameter G
    
95.     % is used to simulate the situation that the chicken swarm have been  
    
96.     % changed, including some chickens have died, or the chicks have grown
    
97.     % up and became roosters or hens, some mother hens have hatched new
    
98.     % offspring (chicks) and so on.
    
99.    
    
100.    if mod( t, G ) == 1 || t == 1   
     
101.                
     
102.         [ ans, sortIndex ] = sort( pFit );   
     
103.         % How the chicken swarm can be divided into groups and the identity
     
104.         % of the chickens (roosters, hens and chicks) can be determined all
     
105.         % depend on the fitness values of the chickens themselves. Hence we
     
106.         % use sortIndex(i) to describe the chicken, not the index i itself.
     
107.         
     
108.         motherLib = randperm( hNum, mNum ) + rNum;   
     
109.         % Randomly select mNum hens which would be the mother hens.
     
110.         % We assume that all roosters are stronger than the hens, likewise,
     
111.         % hens are stronger than the chicks.In CSO, the strong is reflected
     
112.         % by the good fitness value. Here, the optimization problems is
     
113.         % minimal ones, thus the more strong ones correspond to the ones  
     
114.         % with lower fitness values.
     
115.   
     
116.         % Given the fact the 1 : rNum chickens' fitness values maybe not
     
117.         % the best rNum ones. Thus we use sortIndex( 1 : rNum ) to describe
     
118.         % the roosters. In turn, sortIndex( (rNum + 1) :(rNum + 1 + hNum ))
     
119.         % to describle the mother hens, .....chicks.
     
120. 
     
121.         % Here motherLib include all the mother hens.
     
122.       
     
123.        %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
     
124.         % Randomly select each hen's mate, rooster. Hence we can determine
     
125.         % which group each hen inhabits using "mate".Each rooster stands
     
126.         % for a group.For simplicity, we assume that there exist only one
     
127.         % rooster and at least one hen in each group.
     
128.         mate = randpermF( rNum, hNum );
     
129.         
     
130.         % Randomly select cNum chicks' mother hens
     
131.         mother = motherLib( randi( mNum, cNum, 1 ) );  
     
132.    end
     
133.    
     
134.    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%   
     
135.    for i = 1 : rNum                % Update the rNum roosters' values.
     
136.         
     
137.         % randomly select another rooster different from the i (th) one.
     
138.         anotherRooster = randiTabu( 1, rNum, i, 1 );  
     
139.         if( pFit( sortIndex( i ) ) <= pFit( sortIndex( anotherRooster ) ) )
     
140.             tempSigma = 1;
     
141.         else
     
142.             tempSigma = exp( ( pFit( sortIndex( anotherRooster ) ) - ...
     
143.                 pFit( sortIndex( i ) ) ) / ( abs( pFit( sortIndex(i) ) )...
     
144.                 + realmin ) );
     
145.         end
     
146.         
     
147.         x( sortIndex( i ), : ) = pX( sortIndex( i ), : ) .* ( 1 + ...
     
148.             tempSigma .* randn( 1, dim ) );
     
149.         x( sortIndex( i ), : ) = Bounds( x( sortIndex( i ), : ), lb, ub );
     
150.         fit( sortIndex( i ) ) = FitFunc( x( sortIndex( i ), : ) );
     
151.     end
     
152.    
     
153.     for i = ( rNum + 1 ) : ( rNum + hNum )  % Update the hNum hens' values.
     
154.         
     
155.         other = randiTabu( 1,  i,  mate( i - rNum ), 1 );  
     
156.         % randomly select another chicken different from the i (th)  
     
157.         % chicken's mate. Note that the "other" chicken's fitness value  
     
158.         % should be superior to that of the i (th) chicken. This means the   
     
159.         % i (th) chicken may steal the better food found by the "other"
     
160.         % (th) chicken.
     
161.         
     
162.         c1 = exp( ( pFit( sortIndex( i ) ) - pFit( sortIndex( mate( i - ...
     
163.             rNum ) ) ) )/ ( abs( pFit( sortIndex(i) ) ) + realmin ) );
     
164.             
     
165.         c2 = exp( ( -pFit( sortIndex( i ) ) + pFit( sortIndex( other ) )));
     
166. 
     
167.         x( sortIndex( i ), : ) = pX( sortIndex( i ), : ) + ( pX(...
     
168.             sortIndex( mate( i - rNum ) ), : )- pX( sortIndex( i ), : ) )...
     
169.              .* c1 .* rand( 1, dim ) + ( pX( sortIndex( other ), : ) - ...
     
170.              pX( sortIndex( i ), : ) ) .* c2 .* rand( 1, dim );
     
171.         x( sortIndex( i ), : ) = Bounds( x( sortIndex( i ), : ), lb, ub );
     
172.         fit( sortIndex( i ) ) = FitFunc( x( sortIndex( i ), : ) );
     
173.     end
     
174.    
     
175.     for i = ( rNum + hNum + 1 ) : pop    % Update the cNum chicks' values.
     
176.         x( sortIndex( i ), : ) = pX( sortIndex( i ), : ) + ( pX( ...
     
177.             sortIndex( mother( i - rNum - hNum ) ), : ) - ...
     
178.             pX( sortIndex( i ), : ) ) .* FL( i );
     
179.         x( sortIndex( i ), : ) = Bounds( x( sortIndex( i ), : ), lb, ub );
     
180.         fit( sortIndex( i ) ) = FitFunc( x( sortIndex( i ), : ) );
     
181.     end
     
182.    
     
183.     %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
     
184.    % Update the individual's best fitness vlaue and the global best one
     
185.    
     
186.     for i = 1 : pop
     
187.         if ( fit( i ) < pFit( i ) )
     
188.             pFit( i ) = fit( i );
     
189.             pX( i, : ) = x( i, : );
     
190.         end
     
191.         
     
192.         if( pFit( i ) < fMin )
     
193.             fMin = pFit( i );
     
194.             bestX = pX( i, : );
     
195.         end
     
196.     end
     
197. end
     
198. % End of the main program
     
199. 
     
200. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
     
201. % The following functions are associated with the main program
     
202. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
     
203. % This function is the objective function
     
204. function y = Sphere( x )
     
205. y = sum( x .^ 2 );
     
206. 
     
207. % Application of simple limits/bounds
     
208. function s = Bounds( s, Lb, Ub)
     
209.   % Apply the lower bound vector
     
210.   temp = s;
     
211.   I = temp < Lb;
     
212.   temp(I) = Lb(I);
     
213.   
     
214.   % Apply the upper bound vector
     
215.   J = temp > Ub;
     
216.   temp(J) = Ub(J);
     
217.   % Update this new move
     
218.   s = temp;
     
219. 
     
220. %--------------------------------------------------------------------------
     
221. % This function generate "dim" values, all of which are different from
     
222. %  the value of "tabu"
     
223. function value = randiTabu( min, max, tabu, dim )
     
224. value = ones( dim, 1 ) .* max .* 2;
     
225. num = 1;
     
226. while ( num <= dim )
     
227.     temp = randi( [min, max], 1, 1 );
     
228.     if( length( find( value ~= temp ) ) == dim && temp ~= tabu )
     
229.         value( num ) = temp;
     
230.         num = num + 1;
     
231.     end
     
232. end
     
233. 
     
234. %--------------------------------------------------------------------------
     
235. function result = randpermF( range, dim )
     
236. % The original function "randperm" in Matlab is only confined to the
     
237. % situation that dimension is no bigger than dim. This function is
     
238. % applied to solve that situation.
     
239. 
     
240. temp = randperm( range, range );
     
241. temp2 = randi( range, dim, 1 );
     
242. index = randperm( dim, ( dim - range ) );
     
243. result = [ temp, temp2( index )' ];

复制代码

代码下载：https://pan.baidu.com/s/18So1iW4q2ifcMruSuksqEg