We present a suite of fast and effective algorithms, encapsulated in a software package called

Download paper in PDF

We discuss the design, implementation and performance of algorithms suitable for the efficient computation of

Download paper in PDF

A significant number of large optimization problems exhibit structure known as

Download paper in PDF

Many scientific applications benefit from the accurate and efficient computation of derivatives. Automatically generating these derivative computations from an application's source code offers a competitive alternative to other approaches, such as less accurate numerical approximations or labor-intensive analytical implementations. ADIC2 is a source transformation tool for generating code for computing the derivatives (e.g., Jacobian or Hessian) of a function given the C or C++ implementation of that function. Often the Jacobian or Hessian is sparse and presents the opportunity to greatly reduce storage and computational requirements in the automatically generated derivative computation. ColPack is a tool that compresses structurally independent columns of the Jacobian and Hessian matrices through graph coloring approaches. In this paper, we describe the integration of ColPack coloring capabilities into ADIC2, enabling accurate and efficient sparse Jacobian computations. We present performance results for a case study of a simulated moving bed chromatography application. Overall, the computation of the Jacobian by integrating ADIC2 and ColPack is approximately 40% faster than a comparable implementation that integrates ADOL-C and ColPack when the Jacobian is computed multiple times.

Download paper in PDF

The distance-2 graph coloring problem aims at partitioning the vertex set of a graph into the fewest sets consisting of vertices pairwise at distance greater than two from each other. Its applications include derivative computation in numerical optimization and channel assignment in radio networks. We present efficient, distributed-memory, parallel heuristic algorithms for this NP-hard problem as well as for two related problems used in the computation of Jacobians and Hessians. Parallel speedup is achieved through graph partitioning, speculative (iterative) coloring, and a BSP-like organization of parallel computation. Results from experiments conducted on a PC cluster employing up to 96 processors and using large-size real-world as well as synthetically generated test graphs show that the algorithms are scalable. In terms of quality of solution, the algorithms perform remarkably well---the number of colors used by the parallel algorithms was observed to be very close to the number used by the sequential counterparts, which in turn are quite often near optimal. Moreover, the experimental results show that the parallel distance-2 coloring algorithm compares favorably with the alternative approach of solving the distance-2 coloring problem on a graph $\G$ by first constructing the square graph $\G^2$ and then applying a parallel distance-1 coloring algorithm on $\G^2$. Implementations of the algorithms are made available via the Zoltan load-balancing library.

Download paper in PDF

The computation of a sparse Hessian matrix H using automatic differentiation (AD) can be made efficient using the following four-step procedure.

- Determine the
*sparsity structure*of H. - Obtain a
*seed*matrix S that defines a column partition of H using a specialized*coloring*on the adjacency graph of H. - Compute the
*compressed*Hessian matrix B=HS using AD. -
*Recover*the numerical values of the entries of H from B.

Download paper in PDF

A Case Study in a Simulated Moving Bed Process

Using a model from a chromatographic separation process in chemical engineering, we demonstrate that large, sparse Jacobians of fairly complex structures can be computed accurately and efficiently by using automatic differentiation (AD) in combination with a four-step procedure involving matrix

Download paper in PDF

Acyclic and star coloring problems are specialized vertex coloring problems that arise in the efficient computation of Hessians using automatic differentiation or finite differencing, when both sparsity and symmetry are exploited. We present an algorithmic paradigm for finding heuristic solutions for these two NP-hard problems. The underlying common technique is the exploitation of the structure of two-colored induced subgraphs. For a graph

Download paper in PDF

Graph coloring has been employed since the 1980s to efficiently compute sparse Jacobian and Hessian matrices using either finite differences or automatic differentiation. Several coloring problems occur in this context, depending on whether the matrix is a Jacobian or a Hessian, and the specifics of the computational techniques employed. We consider eight variant vertex coloring problems here. This article begins with a gentle introduction to the problem of computing a sparse Jacobian, followed by an overview of the historical development of the research area. Then we present a unifying framework for the graph models of the variant matrix estimation problems. The framework is based upon the viewpoint that a partition of a matrix into structurally orthogonal groups of columns corresponds to distance-2 coloring an appropriate graph representation. The unified framework helps integrate earlier work and leads to fresh insights; enables the design of more efficient algorithms for many problems; leads to new algorithms for others; and eases the task of building graph models for new problems. We report computational results on two of the coloring problems to support our claims. Most of the methods for these problems treat a column or a row of a matrix as an atomic entity, and partition the columns or rows (or both). A brief review of methods that do not fit these criteria is provided. We also discuss results in discrete mathematics and theoretical computer science that intersect with the topics considered here.

Download paper in PDF