Numerical solution of large sets of algebraic nonlinear equations
Mathematics of Computation
Successive column correction algorithms for solving sparse nonlinear systems of equations
Mathematical Programming: Series A and B
A new sparsity preserving quasi-Newton update for solving nonlinear equations
SIAM Journal on Scientific and Statistical Computing
Comparing algorithms for solving sparse nonlinear systems of equations
SIAM Journal on Scientific and Statistical Computing
Representations of quasi-Newton matrices and their use in limited memory methods
Mathematical Programming: Series A and B
Inexact trust region method for large sparse systems of nonlinear equations
Journal of Optimization Theory and Applications
A comparative study on methods for convergence acceleration of iterative vector sequences
Journal of Computational Physics
Alternating multi-step quasi-Newton methods for unconstrained optimization
ICCAM '96 Proceedings of the seventh international congress on Computational and applied mathematics
Testing Unconstrained Optimization Software
ACM Transactions on Mathematical Software (TOMS)
LSQR: An Algorithm for Sparse Linear Equations and Sparse Least Squares
ACM Transactions on Mathematical Software (TOMS)
Practical quasi-Newton methods for solving nonlinear systems
Journal of Computational and Applied Mathematics - Special issue on numerical analysis 2000 Vol. IV: optimization and nonlinear equations
The Secant method for simultaneous nonlinear equations
Communications of the ACM
Superlinear Convergence and Implicit Filtering
SIAM Journal on Optimization
Nonlinear optimization methods for uncalibrated robot visual servoing
ICAI'09 Proceedings of the 10th WSEAS international conference on Automation & information
Bayesian Demand Calibration for Dynamic Traffic Simulations
Transportation Science
Hi-index | 0.00 |
Many complex transportation models can be formulated as fixed-point problems. Typical examples are the equilibrium-like models, motivated by the need to capture the interaction between the transport supply (that is, the infrastructure) and transport demand (that is, travelers behavior) in various ways. In this paper, we propose a new approach for solving fixed-point problems and, equivalently, systems of nonlinear equations. It is a generalization of secant methods, and uses several previous iterates to generate a linear approximation of the nonlinear function. Although it belongs to the quasi-Newton family of methods, our algorithm is matrix free, allowing it to solve large-scale systems of equations without derivative and without any particular assumption about the structure of the problem or its Jacobian. Computational experiments on standard problems show that this algorithm outperforms classical, large-scale quasi-Newton methods in terms of efficiency and robustness. Its numerical performances are similar to Newton-Krylov methods currently considered as the best algorithms to solve large-scale nonlinear systems of equations. Furthermore, our approach, using multiple previous iterates to build a linear approximation of the nonlinear function, exhibits robust behavior in the presence of noise in the equations, which makes it particularly well adapted to transportation problems. We run our algorithm on a simple origin-destination (OD) matrix estimation problem and on some instances of the consistent anticipatory route guidance (CARG) problem, to illustrate the applicability of the proposed method and to show its significant superiority to the classical method of successive averages (MSA). Regarding size, we were able to solve a CARG problem with more than 120,000 variables. We were also able to solve classical problems with up to two million variables.