Searcharxiv⌕ Search

arXiv subjects

David Yang Gao

Publications and source records attributed to David Yang Gao.

At least 19 recordsLinked to original sources
arXiv2018

Canonical Duality Theory and Algorithm for Solving Bilevel Knapsack Problems with Applications

A novel canonical duality theory (CDT) is presented for solving general bilevel mixed integer nonlinear optimization governed by linear and quadratic knapsack problems. It shows that the challenging knapsack problems can be solved analytically in term of their canonical dual solutions. The existence and uniqueness of these analytical solutions are proved. NP-Hardness of the knapsack problems is discussed. A powerful CDT algorithm combined with an alternative iteration and a volume reduction method is proposed for solving the NP-hard bilevel knapsack problems. Application is illustrated by a benchmark problem in optimal topology design. The performance and novelty of the proposed method are compared with the popular commercial codes.

math.OC↗
arXiv2018

A Novel Canonical Duality Theory for Solving 3-D Topology Optimization Problems

This paper demonstrates a mathematically correct and computationally powerful method for solving 3D topology optimization problems. This method is based on canonical duality theory (CDT) developed by Gao in nonconvex mechanics and global optimization. It shows that the so-called NP-hard knapsack problem in topology optimization can be solved deterministically in polynomial time via a canonical penalty-duality (CPD) method to obtain precise 0-1 global optimal solution at each volume evolution. The relation between this CPD method and Gao's pure complementary energy principle is revealed for the first time. A CPD algorithm is proposed for 3-D topology optimization of linear elastic structures. Its novelty is demonstrated by benchmark problems. Results show that without using any artificial technique, the CPD method can provide mechanically sound optimal design, also it is much more powerful than the well-known BESO and SIMP methods. Additionally, computational complexity and conceptual/mathematical mistakes in topology optimization modeling and popular methods are explicitly addressed.

math.OC↗
arXiv2018

On Topology Optimization and Canonical Duality Method

Topology optimization for general materials is correctly formulated as a bi-level knapsack problem, which is considered to be NP-hard in global optimization and computer science. By using canonical duality theory (CDT) developed by the author, the linear knapsack problem can be solved analytically to obtain global optimal solution at each design iteration. Both uniqueness, existence, and NP-hardness are discussed. The novel CDT method for general topology optimization is refined and tested by both 2-D and 3-D benchmark problems. Numerical results show that without using filter and any other artificial technique, the CDT method can produce exactly 0-1 optimal density distribution with almost no checkerboard pattern. Its performance and novelty are compared with the popular SIMP and BESO approaches. Additionally, some mathematical and conceptual mistakes in literature are explicitly pointed out. A brief review on the canonical duality theory for solving a unified problem in multi-scale nonconvex/discrete systems is given in Appendix.

math.OC↗
arXiv2018

On Modelling and Complete Solutions to General Fixpoint Problems in Multi-Scale Systems with Applications

This paper revisits the well-studied fixed point problem from a unified viewpoint of mathematical modeling and canonical duality theory, i.e. the original problem is first reformulated as a nonconvex optimization problem, its well-posedness is discussed based on objectivity principle in continuum physics; then the canonical duality theory is applied for solving this problem to obtain not only all fixed points, but also their stability properties. Applications are illustrated by challenging problems governed by nonconvex polynomial, exponential, and logarithmic operators. This paper shows that within the framework of the canonical duality theory, there is no difference between the fixed point problems and nonconvex analysis/optimization in multidisciplinary studies.

math.OC↗
arXiv2017

Canonical Duality Theory for Topology Optimization

This paper presents a canonical duality approach for solving a general topology optimization problem of nonlinear elastic structures. By using finite element method, this most challenging problem can be formulated as a mixed integer nonlinear programming problem (MINLP), i.e. for a given deformation, the first-level optimization is a typical linear constrained 0-1 programming problem, while for a given structure, the second-level optimization is a general nonlinear continuous minimization problem in computational nonlinear elasticity. It is discovered that for linear elastic structures, first-level optimization is a typical Knapsack problem, which is considered to be NP-complete in computer science. However, by using canonical duality theory, this well-known problem can be solved analytically to obtain exact integer solution. A perturbed canonical dual algorithm (CDT) is proposed and illustrated by benchmark problems in topology optimization. Numerical results show that the proposed CDT method produces desired optimal structure without any gray elements. The checkerboard issue in traditional methods is much reduced.

cs.DM↗
arXiv2016

Multiple nonsmooth solutions for nonconvex variational boundary value problems in $\mathbb{R}^n$

This paper presents a set of complete solutions of a nonconvex variational problem with a double-well potential. Based on the canonical duality-triality theory, the associated nonlinear differential equation with either Dirichlet/Neumann or mixed boundary conditions can be converted into an algebraic equation, which can be solved analytically to obtain all solutions in the dual space. Both global and local extremality conditions are identified by the triality theory. In the application part, typical mechanical models with specificsources and boundary conditions in $\mathbb{R}^2$ are exhibited.

math.OC↗
arXiv2016

On the extrema of a nonconvex functional with double-well potential in higher dimensions

This paper mainly addresses the extrema of a nonconvex functional with double-well potential in higher dimensions through the approach of nonlinear partial differential equations. Based on the canonical duality method, the corresponding Euler--Lagrange equation with Neumann boundary condition can be converted into a cubic dual algebraic equation, which will help find the local extrema for the primal problem. In comparison with the 1D case discussed by D. Gao and R. Ogden, there exists huge difference in higher dimensions, which will be explained in the theorem.

math.OC↗
arXiv2016

On Unified Modeling, Canonical Duality-Triality Theory, Challenges and Breakthrough in Optimization

A unified model is addressed for general optimization problems in multi-scale complex systems. Based on necessary conditions and basic principles in physics, the canonical duality-triality theory is presented in a precise way to include traditional duality theories and popular methods as special applications. Two conjectures on NP-hardness are discussed, which should play important roles for correctly understanding and efficiently solving challenging real-world problems. Applications are illustrated for both nonconvex continuous optimization and mixed integer nonlinear programming. Misunderstandings and confusion on some basic concepts, such as objectivity, nonlinearity, Lagrangian, and Lagrange multiplier method are discussed and classified. Breakthrough from recent false challenges by C. Zălinescu and his co-workers are addressed. This paper will bridge a significant gap between optimization and multi-disciplinary fields of applied math and computational sciences.

math.OC↗
arXiv2015

Global Optimal Trajectory in Chaos and NP-Hardness

This paper presents a new canonical duality methodology for solving general nonlinear dynamical systems. Instead of the conventional iterative methods, the discretized nonlinear system is first formulated as a global optimization problem via the least squares method. The canonical duality theory shows that this nonconvex minimization problem can be solved deterministically in polynomial time if a global optimality condition is satisfied. The so-called pseudo-chaos produced by Runge-Kutta type of linear iterations are mainly due to the intrinsic numerical error accumulations. Otherwise, the global optimization problem could be NP-hard and the nonlinear system can be really chaotic. A conjecture is proposed, which reveals the connection between chaos in nonlinear dynamics and NP-hardness in computer science. The methodology and the conjecture are verified by applications to the well-known logistic equation, a forced memristive circuit and the Lorenz system. Computational results show that the canonical duality theory can be used to identify chaotic systems and to obtain realistic global optimal solutions in nonlinear dynamical systems.

math.OC↗
arXiv2015

Analytic Solutions to Large Deformation Problems Governed by Generalized Neo-Hookean Model

This paper addresses some fundamental issues in nonconvex analysis. By using pure complementary energy principle proposed by the author, a class of fully nonlinear partial diforerential equations in nonlinear elasticity is able to converted a unified algebraic equation, a complete set of analytical solutions are obtained for 3-D finite deformation problems governed by generalized neo-Hookean model. Both global and local extremal solutions to the nonconvex variational problem are identifored by a triality theory. Connection between challenges in nonlinear analysis and NP-hard problems in computational science is revealed. Results show that Legendre-Hadamard condition can only guarantee ellipticity for convex problems. For nonconvex systems, the ellipticity depends not only on the stored energy, but also on the external force field. Uniqueness is proved based on a quasiconvexity and a generalized ellipticity condition. Application is illustrated for logarithm stored energy.

math.AP↗
arXiv2015

On the convexity of nonlinear elastic energies in the right Cauchy-Green tensor

We present a sufficient condition under which a weak solution of the Euler-Lagrange equations in nonlinear elasticity is already a global minimizer of the corresponding elastic energy functional. This criterion is applicable to energies $W(F)=\widehat{W}(F^TF)=\widehat{W}(C)$ which are convex with respect to the right Cauchy-Green tensor $C=F^TF$, where $F$ denotes the gradient of deformation. Examples of such energies exhibiting a blow up for $\det F\to0$ are given.

math.AP↗
arXiv2015

Analytic Solutions to 3-D Finite Deformation Problems Governed by St Venant-Kirchhoff Material

This paper presents a detailed study on analytical solutions to a general nonlinear boundary-value problem in finite deformation theory. Based on canonical duality theory and the associated pure complementary energy principle in nonlinear elasticity proposed by Gao in 1999, we show that the general nonlinear partial differential equation for deformation is actually equivalent to an algebraic (tensor) equation in stress space. For St Venant-Kirchhoff materials, this coupled cubic algebraic equation can be solved principally to obtain all possible solutions. Our results show that for any given external source field such that the statically admissible first Piola-Kirchhoff stress field has no-zero eigenvalues, the problem has a unique global minimal solution, which is corresponding to a positive-definite second Piola-Kirchhoff stress S, and at most eight local solutions corresponding to negative-definite S. Additionally, the problem could have 15 unstable solutions corresponding to indefinite S. This paper demonstrates that the canonical duality theory and the pure complementary energy principle play fundamental roles in nonconvex analysis and finite deformation theory.

math-ph↗
arXiv2014

Double Well Potential Function and Its Optimization in The n-dimensional Real Space -- Part I

A special type of multi-variate polynomial of degree 4, called the double well potential function, is studied. When the function is bounded from below, it has a very unique property that two or more local minimum solutions are separated by one local maximum solution, or one saddle point. Our intension in this paper is to categorize all possible configurations of the double well potential functions mathematically. In part I, we begin the study with deriving the double well potential function from a numerical estimation of the generalized Ginzburg-Landau functional. Then, we solve the global minimum solution from the dual side by introducing a geometrically nonlinear measure which is a type of Cauchy-Green strain. We show that the dual of the dual problem is a linearly constrained convex minimization problem, which is mapped equivalently to a portion of the original double well problem subject to additional linear constraints. Numerical examples are provided to illustrate the important features of the problem and the mapping in between.

math.OC↗
arXiv2014

Canonical Dual Approach for Contact Mechanics Problems with Friction

This paper presents an application of Canonical duality theory to the solution of contact problems with Coulomb friction. The contact problem is formulated as a quasi-variational inequality which solution is found by solving its Karush-Kunt-Tucker system of equations. The complementarity conditions are reformulated by using the Fischer-Burmeister complementarity function, obtaining a non-convex global optimization problem. Then canonical duality theory is applied to reformulate the non-convex global optimization problem and define its optimality conditions, finding a solution of the original quasi-variational inequality. We also propose a methodology for finding the solutions of the new formulation, and report the results on well known instances from literature.

math.OC↗
arXiv2013

Canonical dual solutions to nonconvex radial basis neural network optimization problem

Radial Basis Functions Neural Networks (RBFNNs) are tools widely used in regression problems. One of their principal drawbacks is that the formulation corresponding to the training with the supervision of both the centers and the weights is a highly non-convex optimization problem, which leads to some fundamentally difficulties for traditional optimization theory and methods. This paper presents a generalized canonical duality theory for solving this challenging problem. We demonstrate that by sequential canonical dual transformations, the nonconvex optimization problem of the RBFNN can be reformulated as a canonical dual problem (without duality gap). Both global optimal solution and local extrema can be classified. Several applications to one of the most used Radial Basis Functions, the Gaussian function, are illustrated. Our results show that even for one-dimensional case, the global minimizer of the nonconvex problem may not be the best solution to the RBFNNs, and the canonical dual theory is a promising tool for solving general neural networks training problems.

cs.NE↗
arXiv2012

Canonical Primal-Dual Method for Solving Non-convex Minimization Problems

A new primal-dual algorithm is presented for solving a class of non-convex minimization problems. This algorithm is based on canonical duality theory such that the original non-convex minimization problem is first reformulated as a convex-concave saddle point optimization problem, which is then solved by a quadratically perturbed primal-dual method. %It is proved that the popular SDP method is indeed a special case of the canonical duality theory. Numerical examples are illustrated. Comparing with the existing results, the proposed algorithm can achieve better performance.

math.NA↗
arXiv2012

Complete solutions to nonconvex fractional programming problems

This paper presents a canonical dual approach to the problem of minimizing the sum of a quadratic function and the ratio of nonconvex function and quadratic functions, which is a type of non-convex optimization problem subject to an elliptic constraint. We first relax the fractional structure by introducing a family of parametric subproblems. Under certain conditions, we show that the canonical dual of each subproblem becomes a two-dimensional concave maximization problem that exhibits no duality gap. Since the infimum of the optima of the parameterized subproblems leads to a solution to the original problem, we then derive some optimality conditions and existence conditions for finding a global minimizer of the original problem.

math.OC↗