Archives:Optimization Using Personal Computers
Optimization Using Personal Computers
This practical guide to optimization, or nonlinear programming, provides 33 BASIC computer programs that illustrate the theory and application of methods that automatically adjust design variables. These powerful procedures are available to everyone who uses a personal computer to design or create models in engineering and the sciences. The material emphasizes the interaction between the user and computer by offering hands-on experience with the mathematics and the computational procedures of optimization. It shows how to produce useful answers quickly, while developing a feel for fundamental concepts in matrix algebra, calculus, and nonlinear programming.
Optimization Using Personal Computers reviews the broad range of essential topics of matrix algebra with concrete examples and illustrations, avoiding mathematical abstraction wherever possible. Chapter 1 shows that optimization is intuitively appealing as a geometric interpretation of descent on mathematical surfaces in three dimensions by repetitive computational procedures. Chapter 2 provides a concise review of matrix computations required for optimization. Chapter 3 applies these methods to linear and nonlinear functions of many variables. The three most effective optimization methods are developed, illustrated, and compared in chapters 4 and 5, including nonlinear constraints on the variables. Chapter 6 combines all the best features of the preceding optimization topics with a generally applicable means to compute exact derivatives of responses for networks and their analogues.
This unique book will be of interest to upper-level undergraduates and graduate students, scientists and engineers who use personal computers. These machines have the speed, memory and precision to adjust automatically several dozen variables in complex design problems.
This book provides the programs and theory for these methods and introduces new ways to select and emphasize design objectives, especially for electrical networks and their analogues.