Algorithm 1.

Pseudo-code of CS.

Initialize the fitness value function      f(x),x = (x1,x2,x3,⋯,xd)T      Initialize the number of      iterationst = 0, discovered parameters      Pa = 0.25      Initialize the individual solution $x_i^0$ of      the overall N solutions, (i = 1,2,…,N)      While l ≤ Maximum number of iterations        Using Lévy flight        Update all search agents $x_i^t$ in the        population        Evaluate the new solution xtnew,i for its        fitness value ftnew,i        If the new solution has a better        fitness value          Replace the old solution $x_j^t$ with the          updated solution xtnew,i        End If        Discard some poor-quality solutions        by probability Pa and replace them        with random ones        Increase in the number of iterations        t = t + 1      End While      Return the best solution