Gramacy et al.'s toy problem: minimize x1 + x2 on [0, 1]^2 subject to
  c1 = 1.5 - x1 - 2 x2 - sin(2 pi (x1^2 - 2 x2)) / 2 <= 0, c2 = x1^2 + x2^2 - 1.5 <= 0
global minimum 0.599788 at (0.195123, 0.404665), on the boundary c1 = 0
6 points of a Latin hypercube, then a point per step by log-EI x P(feasible)
evaluation        x1        x2         f         c1         c2  best feasible f - f*
         1  0.193732  0.765545  0.959277  -0.204590  -0.876409               3.6e-1
         2  0.971102  0.545241  1.516343  -0.161850  -0.259672               3.6e-1
         3  0.451976  0.061188  0.513164   0.679542  -1.291974               3.6e-1
         4  0.541108  0.210487  0.751595   0.898457  -1.162897               3.6e-1
         5  0.039182  0.427164  0.466345   0.207200  -1.315996               3.6e-1
         6  0.674825  0.892300  1.567125  -0.520084  -0.248411               3.6e-1
         7  0.000000  0.508147  0.508147   0.534808  -1.241786               3.6e-1
         8  0.000000  0.000000  0.000000   1.500000  -1.500000               3.6e-1
         9  0.000000  0.341614  0.341614   0.360133  -1.383300               3.6e-1
        10  0.000000  0.767732  0.767732  -0.145956  -0.910588               1.7e-1
        11  0.212447  0.400488  0.612935  -0.013086  -1.294476               1.3e-2
        12  0.190897  0.387819  0.578716   0.034616  -1.313155               1.3e-2
        13  0.205325  0.398267  0.603592  -0.001670  -1.299225               3.8e-3
        14  0.191724  0.406987  0.598711   0.001595  -1.297603               3.8e-3
        15  0.193398  0.406780  0.600178  -0.000220  -1.297127               3.9e-4
        16  0.560359  0.000000  0.560359   0.479528  -1.185998               3.9e-4
        17  0.000000  0.201622  0.201622   1.382342  -1.459348               3.9e-4
        18  0.172931  0.425271  0.598202   0.024969  -1.289239               3.9e-4
        19  0.194758  0.405116  0.599874  -0.000089  -1.297951               8.6e-5
        20  0.194735  0.405068  0.599803  -0.000009  -1.297998               1.5e-5
        21  0.194733  0.405064  0.599796  -0.000001  -1.298003               8.4e-6
21 evaluations: the best feasible point (0.194733, 0.405064), 0.599796, 8.4e-6 above the minimum, 5.6e-4 from its point
without the constraints' values: 6 feasible points of 21, the best 0.944530, 3.4e-1 above the minimum
