Data and Feature Reduction in Fuzzy Modeling through Particle Swarm Optimization

Algorithm 1

Pseudocode for cooperative PSO.

Initialize 𝑚 one-dimensional PSO: 𝑃_𝑗, 𝑗∈[1,…,𝑚]

Create

𝐂(𝑗,𝑘)=[𝑃_1(BG),𝑃_2(BG),…,𝑃_𝑗−1(BG),𝑘,𝑃_𝑗+1(BG),…,𝑃_𝑚(BG)]

While stop criteria not met do

    for each subswarm 𝑗∈[1,…,𝑚]  do

      for each particle 𝑖∈[1,…,𝑠]  do

        if fitness(𝐂(𝑗,𝑃_𝑗(𝑥_𝑖))) > fitness(𝐂(𝑗,𝑃_𝑗(LB𝑖)))

           then 𝑃_𝑗(LB𝑖)=𝑃_𝑗(𝑥_𝑖)

        if fitness(𝐂(𝑗,𝑃_𝑗(LB𝑖))) > fitness(𝐂(𝑗,𝑃_𝑗(BG)))

           then  𝑃_𝑗(BG)=𝑃_𝑗(lb𝑖)

   end for

   for each 𝑃_𝑗 do

     𝑣_𝑖,𝑗(𝑡+1)=𝑤⋅𝑣_𝑖,𝑗(𝑡)+𝑐₁⋅𝑟_1,𝑖(𝑡)[𝑃_LB𝑖,𝑗(𝑡)−𝑥_𝑖,𝑗(𝑡)]

           +𝑐₂⋅𝑟_2,𝑖(𝑡)[𝑃_BG𝑖,𝑗(𝑡)−𝑥_𝑖,𝑗(𝑡)]

     𝑥_𝑖,𝑗(𝑡+1)=𝑥_𝑖,𝑗(𝑡)+𝑣_𝑖,𝑗(𝑡+1)