A mixed integer programming approach to the tensor complementarity problem
- 47 Downloads
The tensor complementarity problem is a special instance of nonlinear complementarity problems, which has many applications. How to solve the tensor complementarity problem, via analyzing the structure of the related tensor, is one of very important research issues. In this paper, we propose a mixed integer programming approach for solving the tensor complementarity problem. We reformulate the tensor complementarity problem as an equivalent mixed integer feasibility problem. Based on the reformulation, some conditions for the solution existence and some solution properties of the tensor complementarity problem are given. We also prove that the tensor complementarity problem, corresponding to a positive definite diagonal tensor, has a unique solution. Finally, numerical results are reported to indicate the efficiency of the proposed algorithm.
KeywordsTensor complementarity problem Mixed integer programming Unique solution Positive definite
Mathematics Subject Classification15A69 90C11
The authors would like to thank the editor and the anonymous referees for their helpful constructive comments and suggestions which lead to a significantly improved version of the paper.
- 7.Gowda, M.S., Luo, Z., Qi, L., Xiu, N.: Z-tensors and complementarity problems, arXiv: 1510.07933v2 (2016)
- 8.Huang, Z., Suo, S., Wang, J.: On Q-tensors, arXiv: 1509.03088, (2015)
- 10.Lim, L.-H.: Singular values and eigenvalues of tensors: a variational approach. In: Proceedings of the IEEE International Workshop on Computational Advances in Multi-Sensor Addaptive Processing (CAMSAP’05), Vol. 1, pp. 129–132. IEEE Computer Society Press, Piscataway, NJ (2005)Google Scholar
- 26.Yu, W., Ling, C., He, H.: On the properties of tensor complementarity problems, arXiv:1608.01735v3 (2018)