Efficiently solving linear bilevel programming problems using off-the-shelf optimization software
- 263 Downloads
Many optimization models in engineering are formulated as bilevel problems. Bilevel optimization problems are mathematical programs where a subset of variables is constrained to be an optimal solution of another mathematical program. Due to the lack of optimization software that can directly handle and solve bilevel problems, most existing solution methods reformulate the bilevel problem as a mathematical program with complementarity conditions (MPCC) by replacing the lower-level problem with its necessary and sufficient optimality conditions. MPCCs are single-level non-convex optimization problems that do not satisfy the standard constraint qualifications and therefore, nonlinear solvers may fail to provide even local optimal solutions. In this paper we propose a method that first solves iteratively a set of regularized MPCCs using an off-the-shelf nonlinear solver to find a local optimal solution. Local optimal information is then used to reduce the computational burden of solving the Fortuny-Amat reformulation of the MPCC to global optimality using off-the-shelf mixed-integer solvers. This method is tested using a wide range of randomly generated examples. The results show that our method outperforms existing general-purpose methods in terms of computational burden and global optimality.
KeywordsBilevel programming Mathematical programming with complementarity conditions Nonlinear programming Mixed-integer programming Optimization solvers
This work was supported in part by the Spanish Ministry of Economy, Industry and Competitiveness through Project ENE2016-80638-R and in part by the Research Funding Program for Young Talented Researchers of the University of Málaga through Project PPIT-UMA-B1-2017/18.
- Fletcher R, Leyffer S (2002) Numerical experience with solving MPECs as NLPs. Technical report, Department of Mathematics and Computer Science, University of Dundee, DundeeGoogle Scholar
- Lorenczik S, Malischek R, , Trüby J (2014) Modeling strategic investment decisions in spatial markets. Technical Report 14/09, KölnGoogle Scholar
- Sinha A, Malo P, Deb K (2013) Efficient evolutionary algorithm for single-objective bilevel optimization. arXiv:1303.3901
- The ILOG CPLEX (2015) http://www-01.ibm.com/software/commerce/optimization/cplex-optimizer/index.html
- Von Stackelberg H (1952) The theory of the market economy. Oxford University Press, OxfordGoogle Scholar