HomeConsultingToolboxesProject WorkAbout UsContact Us


Many scientific computing tasks require the repeated computation of derivatives. Hand-coding of derivative functions can be tedious, complex, and error-prone. Moreover, the computation of first and second derivatives, and sometimes the Newton step, is often a dominant part in a scientific computing code. Derivative approximations such as finite-differencing involve additional errors and heuristic choice of parameters.

This toolbox is designed to help a MATLAB user compute first and second derivatives and related structures efficiently, accurately, and automatically. ADMAT 2.0TM employs many sophisticated techniques such as exploiting sparsity and structure to achieve high efficiency in computing derivative structures including gradients, Jacobians, and Hessians. Moreover, ADMAT 2.0 can directly calculate Newton steps for nonlinear systems, often with great efficiency.

A MATLAB user needs only to provide an M-file that evaluates a smooth nonlinear objective function at a given point. On request and when appropriate, ADMAT 2.0 will evaluate the Jacobian matrix (for which the gradient is a special case), the Hessian matrix, and possibly the Newton step in addition to the evaluation of the objective function at the given point. There is no need for the user to provide code for derivative calculation or an approximation scheme.

Please use the following link to download ADMAT 2.0 from us. ADMAT 2.0 is available free of charge for up to one year.

Please click here to obtain a compact guide and a comprehensive guide for using ADMAT 2.0


ADMAT 2.0 includes the following functionalities:

1. Forward mode for first order derivative computation

2. Reverse mode for first order derivative computation

3. Mechanisms and procedures for combining automatic differentiation of M-files with the finite differencing approximation for MEX files

4. Efficient evaluation of sparse Jacobians and sparse Hessians

5. Computation of the sparsity structure of Jacobians and Hessians

6. A template design for the efficient calculation of ‘structured’ Jacobian and Hessian matrices

7. Efficient direct computation of Newton steps while avoiding the full computation of the Jacobian and/or Hessian matrix when appropriate

8. Support of the MATLAB Optimization Toolbox with ADMAT

9. Support of the 1-D MATLAB interpolation function with ADMAT

Please click here for additional information about ADMAT