方法对比
并排查看您选择的方法；存在差异的行会高亮显示。
| 贝叶斯混合整数规划× | 鲁棒混合整数规划× | |
|---|---|---|
| 领域 | 仿真 | 仿真 |
| 方法族 | Process / pipeline | Process / pipeline |
| 起源年份≠ | 2018 (surrogate-BO-MIP synthesis); MIP foundations 1958 | 1998–2004 |
| 提出者≠ | Baptista, R. & Poloczek, M. (formal Bayesian-BO-MIP formulation); mixed-integer programming roots in Gomory (1958) | Ben-Tal & Nemirovski; Bertsimas & Sim |
| 类型≠ | Surrogate-assisted combinatorial optimization | Deterministic robust reformulation of MIP under uncertainty |
| 开创性文献≠ | Baptista, R., Poloczek, M. (2018). Bayesian Optimization of Combinatorial Structures. Proceedings of the 35th International Conference on Machine Learning (ICML), PMLR 80:462–471. link ↗ | Bertsimas, D., Sim, M. (2004). The price of robustness. Operations Research, 52(1), 35–53. DOI ↗ |
| 别名 | Bayesian MIP, BO-MIP, Bayesian Combinatorial Optimization, Mixed-Integer Bayesian Optimization | RMIP, Robust MIP, Uncertain MIP, Robust MILP/MIQP |
| 相关≠ | 5 | 4 |
| 摘要≠ | Bayesian Mixed-Integer Programming (BO-MIP) couples a probabilistic surrogate model — typically a Gaussian process — with a mixed-integer programming solver to efficiently optimize expensive black-box objectives defined over spaces that contain both continuous and discrete or integer-valued decision variables. It is especially valuable when each function evaluation is costly and exhaustive search is infeasible. | Robust Mixed-Integer Programming (RMIP) combines mixed-integer programming with robust optimization to find solutions that remain feasible and near-optimal despite uncertain parameters. Instead of assuming fixed data, it protects decisions against adversarial or worst-case realizations of uncertain inputs, using an explicit uncertainty set to control the degree of conservatism while preserving the combinatorial structure of integer decisions. |
| ScholarGate数据集 ↗ |
|
|