Need precise derivative evaluations? Algorithm 755: ADOL-C presents a robust C++ package designed for automatic differentiation, enabling users to accurately compute first and higher derivatives of vector functions defined within C or C++ programs. This research addresses a crucial need in computational science by providing tools for **derivative evaluation**, **automatic differentiation**, and **C++ package** integration. The core innovation of ADOL-C lies in its ability to evaluate derivatives without truncation errors, ensuring high precision in numerical computations. The resulting derivative routines can be seamlessly integrated into C/C++, Fortran, and other languages. The algorithms efficiently handle derivative matrices by columns or rows, optimizing memory usage and run-time performance. Special routines also cater to solution curves defined by ordinary differential equations, allowing the evaluation of Taylor coefficient vectors and their Jacobians with respect to the current state vector. ADOL-C presents a practical solution for researchers needing accurate derivative calculations, especially in fields like optimization, sensitivity analysis, and scientific computing. Its ability to manage substantial data sequentially, using external files when necessary, further enhances its utility for complex computational tasks.
Published in ACM Transactions on Mathematical Software, this paper presents ADOL-C, a C++ package aligning perfectly with the journal's focus on high-quality mathematical software. The algorithm provides a practical tool for derivative evaluation, a common task in scientific computing. Given the journal's emphasis on software performance and usability, ADOL-C's efficient derivative calculation and easy integration make it a valuable contribution.