A Heuristic for Complementarity Problems Using Difference of Convex Functions

The Linear Dynamic Complementarity Problem is a special case of the Extended Linear Complementarity Problem

In this paper, we consider four optimization models for solving the Linear Complementarity (LCP) Problems. They are all formulated as DC (Difference of Convex functions) programs for which the unified DC programming and DCA (DC Algorithms) are applied. The resulting DCA are simple: they consist of solving either successive linear programs, or successive convex quadratic programs, or simply the projection of points on \(\mathbb{R}_{+}^{2n}\). Numerical experiments on several test problems illustrate the efficiency of the proposed approaches in terms of the quality of the obtained solutions, the speed of convergence, and so on. Moreover, the comparative results with Lemke algorithm, a well known method for the LCP, show that DCA outperforms the Lemke method.

A popular approach for addressing uncertainty in variational inequality problems is by solving the expected residual minimization (ERM) problem. This avenue necessitates distributional information associated with the uncertainty and requires minimizing nonconvex expectation-valued functions. We consider a distinctly different approach in the context of uncertain linear complementarity problems with a view towards obtaining robust solutions. Specifically, we define a robust solution to a complementarity problem as one that minimizes the worst-case of the gap function. In what we believe is amongst the first efforts to comprehensively address such problems in a distribution-free environment, we show that under specified assumptions on the uncertainty sets, the robust solutions to uncertain monotone linear complementarity problem can be tractably obtained through the solution of a single convex program. We also define uncertainty sets that ensure that robust solutions to non-monotone generalizations can also be obtained by solving convex programs. More generally, robust counterparts of uncertain non-monotone LCPs are proven to be low-dimensional nonconvex quadratically constrained quadratic programs. We show that these problems may be globally resolved by customizing an existing branching scheme. We further extend the tractability results to include uncertain affine variational inequality problems defined over uncertain polyhedral sets as well as to hierarchical regimes captured by mathematical programs with uncertain complementarity constraints. Preliminary numerics on uncertain linear complementarity and traffic equilibrium problems suggest that the presented avenues hold promise.

Using a 1958 result of Lojasiewicz, we establish an error bound for analytic systems consisting of equalities and inequalities defined by real analytic functions. In particular, we show that over any bounded region, the distance from any vectorx in the region to the solution set of an analytic system is bounded by a residual function, raised to a certain power, evaluated atx. For quadratic systems satisfying certain nonnegativity assumptions, we show that this exponent is equal to 1/2. We apply the error bounds to the Karush--Kuhn--Tucker system of a variational inequality, the affine variational inequality, the linear and nonlinear complementarity problem, and the 0---1 integer feasibility problem, and obtain new error bound results for these problems. The latter results extend previous work for polynomial systems and explain why a certain square-root term is needed in an error bound for the (monotone) linear complementarity problem.

It is shown that solutions of linear inequalities, linear programs and certain linear complementarity problems (e.g. those with P-matrices or Z-matrices but not semidefinite matrices) are Lipschitz continuous with respect to changes in the right-hand side data of the problem. Solutions of linear programs are not Lipschitz continuous with respect to the coefficients of the objective function. The Lipschitz constant given here is a generalization of the role played by the norm of the inverse of a nonsingular matrix in bounding the perturbation of the solution of a system of equations in terms of a right-hand side perturbation.

The Linear Complementarity Problem LCP(M,q) is to find a vector x in IRn satisfying x0, Mx+q0 and xT(Mx+q)=0, where M as a matrix and q as a vector, are given data. In this paper we show that the linear complementarity problem is completely equivalent to finding the fixed point of the map x = max (0, (I-M)x-q); to find an approximation solution to the second problem, we propose an algorithm starting from any interval vector X(0) and generating a sequence of the interval vector (X(k))k=1 which converges to the exact solution of our linear complementarity problem. We close our paper with some examples which illustrate our theoretical results.

In this paper, we introduce total dual integrality of the linear complementarity problem (LCP) by analogy with the linear programming problem. The main idea of defining the notion is to propose the LCP with orientation, a variant of the LCP whose feasible complementary cones are specified by an additional input vector. Then we naturally define the dual problem of the LCP with orientation and total dual integrality of the LCP. We show that if the LCP is totally dual integral, then all basic solutions are integral. If the input matrix is sufficient or rank-symmetric, and the LCP is totally dual integral, then our result implies that the LCP always has an integral solution whenever it has a solution. We also introduce a class of matrices such that any LCP instance having the matrix as a coefficient matrix is totally dual integral. We investigate relationships between matrix classes in the LCP literature such as principally unimodular matrices. Principally unimodular matrices are known that all basic solutions to the LCP are integral for any integral input vector. In addition, we show that it is coNP-hard to decide whether a given LCP instance is totally dual integral.

This paper is concerned with the mixed linear complementarity problem and the role it and its variants play in the stability analysis of the nonlinear complementarity problem and the Karush-Kuhn-Tucker system of a variational inequality problem. Under a nonsingular assumption, the mixed linear complementarity problem can be converted to the standard problem; in this case, the rich theory of the latter can be directly applied to the former. In this work, we employ degree theory to derive some sufficient conditions for the existence of a solution to the mixed linear complementarity problem in the absence of the nonsingularity property. Next, we extend this existence theory to the mixed nonlinear complementarity problem and establish a main stability result under a certain degree-theoretic assumption concerning the linearized problem. We then specialize this stability result and its consequences to the parametric variational inequality problem under the assumption of a unique set of multipliers. Finally, we consider the latter problem with the uniqueness assumption of the multipliers replaced by a convexity assumption and obtain stability results under some weak second-order conditions. In addition to the new existence results for the mixed linear complementarity problem, the main contributions of this paper in the stability category are the following: a resolution to a conjecture concerning the local solvability of a parametric variational inequality, the use of the generalized linear complementarity problem as a tool to broaden the second-order conditions, the characterization of the solution stability of the linear complementarity problem and the affine variational inequality problem in terms of the solution isolatedness under some weak hypotheses, and various stability theorems under some weak second-order conditions.

The continuous network design problem (CNDP) has been recognized as one of the most challenging issues in the field of transportation. Existing approaches to solving the CNDP are primarily heuristic without convergence guarantee or suitable for handling small networks because of the inherent nonconvexity arising from its bilevel hierarchical structure. An efficient and convergent approach for solving the CNDP on large networks has been fervently sought. In this paper, we present a novel convergent approach centered around exploiting the inherent convexity-related structure within the CNDP. We first demonstrate that the CNDP can be equivalently formulated as a difference of convex (DC) program with all involved functions being either convex functions or DC functions. Then, by exploiting the DC structure, we give a convex programming approximation for the CNDP and subsequently propose a penalized sequential convex programming approach. Finally, we show that the proposed method can yield an approximately stationary point under commonly used conditions. A numerical study is conducted on some real networks from a reputable network repository for transportation research. The numerical results demonstrate that the proposed method achieves better solutions with faster computational speed, particularly on larger networks, as compared with two heuristic approaches and two convergent approaches. History: Accepted by Russell Bent, Area Editor for Network Optimization: Algorithms and Applications. Funding: This research was supported by the National Science Foundation of China [Grants 72131007, 72140006, 12271161, 12222106, and 12326605], the Natural Science Foundation of Shanghai [Grant 22ZR1415900], and Guangdong Basic and Applied Basic Research Foundation [Grant 2022B1515020082]. Supplemental Material: The software that supports the findings of this study is available within the paper and its Supplemental Information ( https://pubsonline.informs.org/doi/suppl/10.1287/ijoc.2024.0737 ) as well as from the IJOC GitHub software repository ( https://github.com/INFORMSJoC/2024.0737 ). The complete IJOC Software and Data Repository is available at https://informsjoc.github.io/ .

Proposed is a time-marching al- gorithm to solve a nonlinear system of com- plementarity equations: Pi(xj) ≥ 0, Qi(xj) ≥ 0, Pi(xj)Qi(xj )= 0, i, j = 1,...,n, resulting from a discretization of nonlinear obstacle problem. We transform the above nonlinear complemen- tarity problem (NCP) into a nonlinear algebraic equations (NAEs) system: Fi(xj )= 0 with the aid of the Fischer-Burmeister NCP-function. Such NAEs are semi-smooth,highlynonlinearand usu- ally implicit, being hard to handle by the Newton- like method. Instead of, a first-order system of ODEs is derived through a fictitious time equa- tion. The time-stepping equations are obtained by applying a numerical integration on the resul- tant ODEs, which are derivative-free and do not need the inverse of any matrix. The computa- tional cost is thus greatly reduced. The numerical examples of Bratu, von Karman and other elliptic equationsareusedto demonstratethatthenewfic- titious time integration method (FTIM) is highly efficient to calculate the obstacle problems. Keyword: Nonlinear Obstacle Problem, Non- linear complementarity problem, Nonlinear alge- braic equations, Iterative method, Elliptic equa- tions, Fictitious time integration method (FTIM)

The boosted difference of convex functions algorithm (BDCA) was recently proposed for minimizing smooth difference of convex (DC) functions. BDCA accelerates the convergence of the classical difference of convex functions algorithm (DCA) thanks to an additional line search step. The purpose of this paper is twofold. First, we show that this scheme can be generalized and successfully applied to certain types of nonsmooth DC functions, namely, those that can be expressed as the difference of a smooth function and a possibly nonsmooth one. Second, we show that there is complete freedom in the choice of the trial step size for the line search, which is something that can further improve its performance. We prove that any limit point of the BDCA iterative sequence is a critical point of the problem under consideration and that the corresponding objective value is monotonically decreasing and convergent. The global convergence and convergence rate of the iterations are obtained under the Kurdyka--Łojasiewicz property. Applications and numerical experiments for two problems in data science are presented, demonstrating that BDCA outperforms DCA. Specifically, for the minimum sum-of-squares clustering problem, BDCA was on average 16 times faster than DCA, and for the multidimensional scaling problem, BDCA was 3 times faster than DCA.

For a given function f: ℝ k → ℝ k we mean by the complementarity problem, corresponding to f, the problem CP: Find x ∈ ℝ k , such that x ≥ 0, f(x) ≥ 0, xT f(x) = 0. If f is an affine function, say f(x) = q + Mx for all x ∈ ℝ n for some n × n-matrix M and q ∈ ℝ n , then we call the problem a linear complementarity problem (LCP). Let us denote the solution set of CP by O(f) and the solution set of LCP, corresponding to (q, M), by O(q, M). Hence, \(O(f) = \{ \hat x \in {\mathbb{R}^k}|\hat x \geqslant 0,f(\hat x) \geqslant 0,{\hat x^ \top }f(\hat x) = 0\}\) etc. Many problems in Operations Research, mathematical economics and game theory can be translated into a complementarity problem. We give some examples in the following exercises.

In this paper we consider a class of structured nonsmooth difference-of-convex (DC) minimization in which the first convex component is the sum of a smooth and nonsmooth functions while the second convex component is the supremum of possibly infinitely many convex smooth functions. We first propose an inexact enhanced DC algorithm for solving this problem in which the second convex component is the supremum of finitely many convex smooth functions, and show that every accumulation point of the generated sequence is an $$(\alpha ,\eta )$$ -D-stationary point of the problem, which is generally stronger than an ordinary D-stationary point. In addition, inspired by the recent work (Pang et al. in Math Oper Res 42(1):95–118, 2017; Wen et al. in Comput Optim Appl 69(2):297–324, 2018), we propose two proximal DC algorithms with extrapolation for solving this problem. We show that every accumulation point of the solution sequence generated by them is an $$(\alpha ,\eta )$$ -D-stationary point of the problem, and establish the convergence of the entire sequence under some suitable assumption. We also introduce a concept of approximate $$(\alpha ,\eta )$$ -D-stationary point and derive iteration complexity of the proposed algorithms for finding an approximate $$(\alpha ,\eta )$$ -D-stationary point. In contrast with the DC algorithm (Pang et al. 2017), our proximal DC algorithms have much simpler subproblems and also incorporate the extrapolation for possible acceleration. Moreover, one of our proximal DC algorithms is potentially applicable to the DC problem in which the second convex component is the supremum of infinitely many convex smooth functions. In addition, our algorithms have stronger convergence results than the proximal DC algorithm in Wen et al. (2018).

A number of equivalent characterizations for the existence and boundedness of solutions of the linear complementarity problem: Mx+q≥0, x≥0, x T(Mx+q)=0 where M is an n×n real matrix and q is an n-vector, are given for the case when M is copositive plus. The special case when M is skew-symmetric covers the linear programming case. One useful characterization of existence and boundedness of solutions is given by solving a simple linear program. Other important characterizations are the Slater constraint qualification and the stability condition that for all arbitrary but sufficiently small perturbations of the data M and q which maintain copositivity plus, the perturbed linear complementarity problem is solvable and its solutions are uniformly bounded. An interesting sufficient condition for boundedness of solutions is that the linear complementarity problem have a nondegenerate vertex solution. Another result is that the subclass ™ of copositive plus matrices for which the linear complementarity problem has a solution for each q in R n, that is ™⊂Q, coincides with the subclass of copositive plus matrices for which the linear complementarity problem has a nonempty bounded solution set for each q in R n.Key wordsLinear Complementarity ProblemSolution SetBoundednessCopositive-plus Matrices

Based on a smoothing symmetric disturbance FB-function, a smoothing inexact Newton method for solving the nonlinear complementarity problem with P0-function was proposed. It was proved that under mild conditions, the given algorithm performed global and superlinear convergence without strict complementarity. For the same linear complementarity problem(LCP), the algorithm needs similar iteration times to the literature. However, its accuracy is improved by at least 4 orders with calculation time reduced by almost 50%, and the iterative number is insensitive to the size of the LCP. Moreover, fewer iterations and shorter time are required for solving the problem by using inexact Newton methods for different initial points.