Algorithm 3. IFPASO algorithm.

1: Start 2: Randomly initialize particle swarm 3: While (number of iterations or the stopping iteration on is not met) 4: Evaluate fitness of the particle swarm 5: for n = 1 to number of particles 6: Find individual optimal solution $p_t^i$ 7: Find group optimal solution $p_t^g$ 8: for d = 1 to number of dimensions of particle 9: update the velocity of particles via $v_{t+1}^i=w_t{v}_t^i+c_1{r}_1(p_t^i-x_t^i)+c_2{r}_2(p_t^g-x_t^i)$ 10: update the position of particles via $x_{t+1}^i=x_t^i+v_{t+1}^i$ 11: end for 12: end for 13: update the inertia weight via $w_t=(w_{\max }-w_{\min })\ast \frac{(t_{\max }-t)}{t_{\max }}+w_{\min }$ 14: end While 15: Output the best solution found 16: The best solution found by PSO is regarded as initial points for FPA algorithm $g_{best}^{*}$ 17: While $t<Max\_generation$ 18: for i=1:n (each flower in the population) 19: get dynamic switch probability $p$ via $0.8-0.1\ast \frac{Max\_T-t}{Max\_T}$ 20: if (rand < p) 21: Draw a (d-dimensional) step vector L which obeys a Lévy distribution 22: Global pollination via $x_i^{t+1}=x_i^t+\gamma L\left(\lambda \right)(x_i^t-g_{best}^{*})$ 23: else 24: Draw $\epsilon$ from a uniform distribution in [0, 1] 25: Do local pollination via $x_i^{t+1}=x_i^t+\epsilon (x_j^t-x_k^t)$ 26: end if 27: Evaluate each new solution $x_i^t$ 28: If new solution is better, update it in the population 29: end for 30: Find the current best solution $g_{best}^{*}$ 31: end While 32: Output the final best solution found