Research Article
The Distributionally Robust Optimization Reformulation for Stochastic Complementarity Problems
Table 1
Numerical comparison between the DROR and the ERM methods.
| | Case | Method | xsol | compl | feas | pr (%) | compl | feas | pr (%) | | Uniform samples | Normal samples |
| | 1 | DROR | (0.000003, 1.000001, 1.000000) | 5.44e − 06 | 1.73e − 07 | 100.0 | 5.76e − 06 | 1.24e − 02 | 96.4 | | DROR | (0.000021, 1.000006, 0.999999) | 3.39e − 05 | 3.06e − 06 | 100.0 | 3.60e − 05 | 1.24e − 02 | 92.9 | | DROR | (0.000001, 1.000000, 1.000000) | 1.44e − 06 | 5.21e − 07 | 100.0 | 1.54e − 06 | 1.24e − 02 | 96.4 | | ERM | (0.000000, 0.999997, 1.000001) | 1.44e − 06 | 5.48e − 06 | 100.0 | 1.53e − 06 | 1.24e − 02 | 96.4 | | ERM | (0.044313, 0.998574, 1.002037) | 7.11e − 02 | 2.49e − 02 | 0.0 | 7.54e − 02 | 3.34e − 02 | 1.8 |
| | 2 | DROR | (6.001528, 1.000509, 7.001366) | 1.81e + 02 | 0.00e + 00 | 100.0 | 1.80e + 02 | 2.10e − 03 | 99.8 | | DROR | (3.001060, 1.000530, 4.000647) | 5.43e + 01 | 0.00e + 00 | 100.0 | 5.40e + 01 | 2.65e − 02 | 97.4 | | DROR | (1.583408, 1.000545, 2.582745) | 1.97e + 01 | 2.05e − 02 | 91.2 | 1.96e + 01 | 8.26e − 02 | 89.3 | | ERM | (0.367029, 0.911208, 1.082520) | 1.65e + 00 | 3.09e − 01 | 27.5 | 1.60e + 00 | 3.48e − 01 | 22.4 | | ERM | (0.403395, 0.841871, 1.130820) | 1.92e + 00 | 3.84e − 01 | 15.3 | 1.88e + 00 | 4.38e − 01 | 9.6 |
|
|
