Some of the most fundamental problems in engineering, science, and mathematics would take the most powerful computer in the world several lifetimes to find an optimal solution. However, near-optimal solutions to many of these problems have been discovered thanks to various methods of mathematical optimization. This Topic in Depth looks at some optimization techniques and the areas to which they have been applied.
Argonne National Laboratory hosts an online guide (1) to some of the most well known optimization problems and algorithms. People who are new to the subject can find a wealth of introductory material in the Optimization Tree section, and several applications are illustrated with interactive demonstrations in the Case Studies section. Optimization has roots in operations research, and this tutorial (2) covers many topics within OR. A fun applet requires the user to place as many queens on a chess board as possible without any two being in direct line-of-sight, and an accompanying discussion shows how linear programming can be used to solve this problem; this is one of many resources contained within the tutorial. Two chemical engineering professors at Carnegie Mellon University are the authors of Retrospective on Optimization (3), a fairly comprehensive paper chronicling the history of optimization problems and the development of solution methods. The 51-page document is divided into two main parts; the first outlines some of the most significant advances in the field, and the second looks ahead toward key areas of research needed to evolve optimization further. Highway planning and development is the focus of this paper (4), which proposes using, among other things, genetic algorithms to optimize highway alignment. The authors state that this technique could be used to avoid delays and added costs due to changing plans later in the construction process. Another use of genetic algorithms in optimization is highlighted in this document from the German Aerospace Center (5). The document describes how a novel method for robot design, which involves formulating mathematical representations of robotic constraints and kinematics. These figures can then be used as parameters in a genetic algorithm that would optimize component placement while maintaining the desired functionality. A joint effort between architecture and mechanical engineering researchers at the University of Michigan (6) applied geometrical and topological optimization techniques to building floorplan layout. In the course of their research, they developed an optimization tool that is briefly described and can be downloaded from a Web site given in the paper. The Mathematical Programming Glossary (7) contains short discussions of hundreds of terms related to optimization theory. Each entry includes a clear definition and hyperlinks to other terms, and some of the entries are accompanied by a more in-depth supplement. One of the best sources for current research papers and reports is Optimization Online (8). This repository receives numerous submissions each month, and papers can be browsed by data added or category.