A single-level reformulation of mixed integer bilevel programming problems
Introduction
In this paper, we consider the mixed integer bilevel programming problem (MIBLPP) of the form where some of the upper-level decision variables are integer-valued, , is a nonempty closed subset in for any , and is the set of global minimizers for the lower-level problem parameterized by where represents the lower-level continuous variables, represents the lower-level integer variables, is a continuously differentiable function with respect to , and is defined as with a set of integers in and continuously differentiable functions with respect to . Throughout this paper, we assume that the cardinality of is finite and .
Bilevel programming problems have been intensively investigated; see, e.g., the review articles [6], [7]. Almost all investigations in the literature are devoted to problems with only continuous variables in the lower-level problem for which the popular way is to reformulate it as a single-level problem, although there exist many practical problems with integer variables in the lower-level problem; see, e.g., [2], [4], [9], [18]. There have also been some progresses made on the numerical methods for solving linear MIBLPPs; see, e.g, [2], [3], [17], [20], [23], [24].
It is desirable to know whether there exists a minimizer before solving the MIBLPP. Vicente et al. [22] studied the existence of minimizers of linear MIBLPP for different cases corresponding to particularizations of the upper-level and lower-level variables. Unfortunately, for the case where there exist lower-level integer variables and upper-level joint constraints, it is a difficult task to give the existence of minimizers. In this paper, we assume that a minimizer exists for the MIBLPP and focus on deriving its single-level reformulation.
To the best of our knowledge, there are only a few publications on developing numerical methods for solving MIBLPPs with nonlinear functions [10], [12], [16]. Our proposed single-level reformulation in this paper makes it possible to develop more approaches for solving MIBLPPs as what has been done for bilevel programming problems with only continuous variables in the lower-level problem.
When there is no lower-level integer variable (), we may replace the lower-level problem of the MIBLPP with its Karush–Kuhn–Tucker (KKT) conditions, resulting in a mathematical program with complementarity constraints (MPCC). Their relations in the sense of global and local minimizers have been established in [8] provided that the lower-level problem is convex and Slater’s constraint qualification is satisfied. When there are lower-level integer variables (), the lower-level problem is clearly nonconvex and thus we cannot replace the lower-level problem with its KKT conditions directly. In this paper, we reformulate the MIBLPP as a mixed integer mathematical program with complementarity constraints (MIMPCC) by separating the lower-level continuous and integer variables. In order to investigate the relations between the MIBLPP and the MIMPCC, we assume that for any lower-level integer variables, the lower-level problem is solvable and convex with respect to the continuous variables. We show that the global minimizers of the MIBLPP correspond to those of the MIMPCC provided that for any lower-level integer variables, the lower-level problem satisfies the generalized Slater’s constraint qualification with respect to the continuous variables. Since bilevel programming is generally nonconvex, it is difficult to find a global minimizer and in some cases, one needs to be happy with obtaining a local minimizer. Thus, it is also necessary to investigate the relations between the local minimizers of the MIBLPP and the MIMPCC. The stronger Slater’s constraint qualification and the so-called restricted inf-compactness condition are required to ensure that the local minimizers of the MIBLPP correspond to those of the MIMPCC.
The rest of this paper is organized as follows. In Section 2 we give some background materials and preliminary results. In Section 3 we reformulate the MIBLPP as a single-level MIMPCC. In particular, we show that the global and local minimizers of the MIBLPP correspond to those of the MIMPCC respectively under some suitable conditions. Section 4 concludes the paper.
Notation: For any vectors , means that vector and vector are perpendicular. Given a point and a function , denotes the transposed Jacobian of at .
Section snippets
Preliminaries and preliminary results
We first review some constraint qualifications for convex constrained sets. Let be a convex constrained set defined by with a convex and continuously differentiable function and a linear function . The following two constraint qualifications will be useful in the next section. Definition 1 We say that Slater’s constraint qualification (SCQ) holds for if has the full column rank and there exists such that We say that the generalized Slater’s
Single-level reformulation
This section focuses on the single-level reformulation of the MIBLPP. Let Assumption 1 hold throughout this section. Assumption 1 Assume that for any , and are convex and continuously differentiable functions, and is a linear function.
In order to give a single-level reformulation of the MIBLPP, we need use a system of equalities and inequalities to characterize the optimal solution set of problem (2). Although problem (2) is a nonconvex problem, we observe that for
Concluding remarks
We have shown that the MIBLPP may be reformulated as a single-level MIMPCC by introducing variables and constraints. If the number of lower-level integer variables is very small, the MIMPCC will be relatively easy to solve. If is large, the MIMPCC will be large-scaled and then it is also very difficult to solve. In this case, the column-and-constraint generation algorithm proposed in [24] may be an efficient approach to solve the MIMPCC. We leave the challenging problem
Acknowledgments
The authors are grateful to the referees for their helpful suggestions and comments. This work was supported in part by the NSFC (No. 11401379).
References (24)
- et al.
One-level reformulation of the bilevel knapsack problem using dynamic programming
Discrete Optim.
(2013) - et al.
Discrete bilevel programming: Application to a natural gas cash-out problem
European J. Oper. Res.
(2005) - et al.
A bilevel formulation of the pediatric vaccine pricing problem
European J. Oper. Res.
(2016) - et al.
An exact algorithm for the bilevel mixed integer linear programming problem under three simplifying assumptions
Comput. Oper. Res.
(2014) - et al.
Links between linear bilevel and mixed 0–1 programming problems
J. Optim. Theory Appl.
(1997) - et al.
An algorithm for the discrete bilevel programming problem
Naval Res. Logist.
(1992) - et al.
Optimal pricing and rebate strategies in a two-level supply chain
Prod. Oper. Manage.
(2009) Optimization and Nonsmooth Analysis
(1983)- et al.
An overview of bilevel optimization
Ann. Oper. Res.
(2007) Annotated bibliography on bilevel programming and mathematical programs with equilibrium constraints
Optimization
(2003)
Is bilevel programming a special case of a mathematical program with complementarity constraints?
Math. Program.
Bilevel programming with discrete lower level problems
Optimization
Cited by (10)
Bilevel aggregator-prosumers' optimization problem in real-time: A convex optimization approach
2022, Operations Research LettersCitation Excerpt :This price oriented setup falls into the category of bilevel optimization problems [5] and Stackelberg games [23], where the lower level problem and the upper level problem are the problems related to the prosumers and the aggregator, respectively. Bilevel optimization approach has been used in various applications with hierarchical nature [8,13,14]. Particularly, it has extensively used to model and solve energy systems problems [7].
Mathematical programs with multiobjective generalized Nash equilibrium problems in the constraints
2021, Operations Research LettersCitation Excerpt :In this paper, we call such kind of problems “the mathematical program with multiobjective generalized Nash equilibrium problems in the constraints”. One main approach to deal with problems with hierarchical structure is to transform them into single-level optimization problems by replacing the lower-level problem with its KKT conditions, and equivalence results are given in the sense of global and local minimizers under some mild conditions (see, e.g., [17,30]). The transformed reformulation is a special case of the so-called mathematical program with equilibrium constraints (MPEC).
Integrating link-based discrete credit charging scheme into discrete network design problem
2019, European Journal of Operational ResearchCitation Excerpt :Later, Mitsos (2010) extended the deterministic algorithm for the global solution of continuous bilevel programs with a nonconvex lower-level program to the mixed-inter case. Recently, Li and Guo (2017) proposed the algorithm for the mixed integer bilevel programming problem via reformulating it as a mixed integer mathematical program with complementarity constraints. However, it notes that no rigorous approaches are available for solving bilevel mixed-integer programming problems.
Mixed integer semi-infinite programming problem
2021, Scientia Sinica MathematicaThe development of optimization and decision theory in systems engineering
2020, Xitong Gongcheng Lilun yu Shijian/System Engineering Theory and PracticeBilevel Optimization: Theory, Algorithms, Applications and a Bibliography
2020, Springer Optimization and Its Applications