Rosenbrock's function in [-2, 0.5] x [-1, 3], from (-1.2, 1); its minimum (1, 1) is outside
L-BFGS-B, the current point after each round
round  evaluations         x1         x2          f  projected gradient
    0            1  -1.200000   1.000000  24.200000               2.0e0
    1            2  -1.200000   1.000000  24.200000               2.0e0
    2            3  -1.068165   1.155100   4.297256               2.2e0
    3            4  -1.069587   1.150927   4.287965               1.4e0
    4            5  -1.069176   1.147217   4.283154               1.6e0
    5            6  -1.064751   1.130668   4.264113               1.6e0
    6            7  -1.050615   1.090760   4.222010               1.9e0
    7            8  -1.010282   0.991250   4.127783               2.0e0
    8            9  -0.947934   0.853665   3.996166               2.1e0
    9           10  -0.871791   0.708197   3.772158               2.3e0
   10           11  -0.750285   0.519042   3.256083               2.5e0
   11           12  -0.643758   0.404061   2.712679               2.1e0
   12           13  -0.643758   0.404061   2.712679               2.1e0
   13           14  -0.549242   0.269766   2.501914               2.7e0
   14           15  -0.382191   0.085933   2.272099               2.9e0
   15           16  -0.357501   0.115638   1.857615               2.4e0
   16           17  -0.202458   0.013388   1.522091               3.0e0
   17           18  -0.054524  -0.045690   1.348826               3.0e0
   18           19  -0.001524  -0.002038   1.003466              5.0e-1
   19           20   0.171300  -0.010348   0.844285               3.0e0
   20           21   0.151224   0.014616   0.727231               1.7e0
   21           22   0.245888   0.051377   0.576937               1.8e0
   22           23   0.245888   0.051377   0.576937               1.8e0
   23           24   0.328856   0.087658   0.492410               2.9e0
   24           25   0.434179   0.162541   0.387599               2.8e0
   25           26   0.495001   0.244737   0.255032              5.8e-2
   26           27   0.500000   0.249614   0.250015              7.7e-2
   27           28   0.500000   0.250040   0.250000              8.0e-3
   28           29   0.500000   0.250000   0.250000               0.0e0
L-BFGS-B: converged at (0.5, 0.25), f = 0.25, after 29 evaluations
the gradient there is (-1.0, 0.0): f falls only beyond the bound x1 = 0.5
Nelder-Mead, for contrast: (0.49999999999999956, 0.24999999980867693), f = 0.25000000000000044, after 257 evaluations
