Ant Colony Optimization with Three Stages for Independent Test Cost Attribute Reduction

Algorithm 1

ACO with distributed deletion to MTR problem.

Input: 𝑆=(𝑈,𝐶,𝐷,𝑉,𝐼,𝑐)

Output: 𝑀, a reduct of 𝑆

Method: acomtr

(1) 𝑀=∅;//the best reduct obtained by ants

(2) for (𝑖=0; 𝑖<𝐾; 𝑖 ++) do

(3)  for (𝑗=0; 𝑗<𝑘; 𝑗++) do

(4)    create ant[𝑖×𝑘+𝑗];

(5)   𝐵_{𝑖×𝑘+𝑗}=core(𝑆);//takes the set of core attributes

(6)  end for

  //Each ant selects attributes

(7)  for (each 𝑚∈[𝑖×𝑘,(𝑖+1)×𝑘] where ant[𝑚] has not stopped) do

(8)   MonteCarlo(𝑡𝑐,𝑝ℎ);//select the next vertex

(9)   if (𝑃𝑂𝑆_𝐵𝑚(𝐷)==𝑃𝑂𝑆_𝐶(𝐷)) then

(10)    ant[𝑚] stops; //Delete redundant vertexes.

(11)    𝑚𝑡𝑐=10000;//minimal test cost

(12)    for (𝑗=|𝐵_𝑖|−1; 𝑗>=|Core(𝑆)|; 𝑗- -) do

(13)     if (𝑃𝑂𝑆_{(𝐵𝑖−{𝑎𝑗})}(𝐷)==𝑃𝑂𝑆_𝐵𝑖(𝐷)) then

(14)      𝐵_𝑖=𝐵_𝑖−{𝑎_𝑗};

(15)     end if

(16)    end for

(17)    Adjust the traveled path

(18)    Update the pheromone of edges ant[𝑚] traveled

(19)   end if

(20)  end for

(21) end for

 //Choose the minimal test cost reduction from the last 𝑛 ants

(22) for (𝑖=𝐾×𝑘−𝑛; 𝑖<=𝐾×𝑘−1; 𝑖 ++) do

(23)    if  𝑡𝑐(𝐵_𝑖)≤𝑚𝑡𝑐  then

(24)    𝑚𝑡𝑐=𝑡𝑐(𝐵_𝑖);

(25)   𝑀=𝐵_𝑖;

(26)  end if

(27) end for

(28) return 𝑀;