xavier-mayiming / enhanced-grey-wolf-optimizer Goto Github PK

$$ \begin{align} &\text{min}\ f(x)=0.6224x_1x_3x_4+1.7781x_2x_3^2+3.1661x_1^2x_4+19.84x_1^2x_3,\\ &\text{s.t.} \\ &-x_1+0.0193x_3\leq0,\\ &-x_3+0.0095x_3\leq0,\\ &-\pi x_3^2x_4-\frac{4}{3}\pi x_3^3+1296000\leq0,\\ &x_4-240\leq0,\\ &0\leq x_1\leq99,\\ &0\leq x_2 \leq99,\\ &10\leq x_3 \leq 200,\\ &10\leq x_4 \leq 200. \end{align} $$

if __name__ == '__main__':
    # Parameter settings
    pop = 50
    iter = 2000
    lb = [0, 0, 10, 10]
    ub = [99, 99, 200, 200]
    print(main(pop, iter, lb, ub))

The EGWO converges at its 1,271-th iteration, and the global best value is 8050.913534658795.

{
    'best score': 8050.913534658795, 
    'best solution': [1.3005502034963052, 0.6428626394484327, 67.3860209065443, 10.0], 
    'convergence iteration': 1271
}

The GWO code: https://github.com/Xavier-MaYiMing/Grey-Wolf-Optimizer

$$ f(x)=\sum_{i=1}^{30}(x_i-0.0001)^2,\qquad x_i\in[-10, 100], \quad i=1,\cdots,30. $$

$$ f(x)=\sum_{i=1}^{30}((x_i-1)^2-10\cos(2\pi(x_i-1)) + 10),\qquad x_i\in[-4.12, 5.12], \quad i=1,\cdots,30. $$

If the global optimal solution shifts away from the origin of the coordination, the performance of the original GWO deteriorates sharply. The EGWO solves this problem, and its performance on shifted functions proves its effectiveness.