
Accession Number : ADA299010
Title : Optimization of Linearly Constrained Indefinite Functions,
Corporate Author : NAVAL ACADEMY ANNAPOLIS MD
Personal Author(s) : Hildebrandt, Wesley A.
PDF Url : ADA299010
Report Date : 09 MAY 1995
Pagination or Media Count : 85
Abstract : Global optimization is the process of finding a best solution among many possible solutions to a problem involving the minimization or maximization of some desired "cost" function. In general, many problems arising from practical applications can be formulated using both an objective function to be optimized (cost, profit, etc.) and a set of restrictions on the allowed solutions. In some cases this objective function may be linear, in which case the problem may yield to linear programming techniques. In other cases it may be entirely concave or convex. In these cases the solution may again be easy to obtain since certain properties of the function allow special searching techniques to locate the optimum solution. In the hardest cases the objective function is indefinite, which means that it can have many local minima, none of which satisfy any special properties. Furthermore, these problems are usually bounded by constraints, which restrict the allowed values of the individual variables. In the majority of real problems the constraints will be linear in which case the optimization problem can be approached using matrix algebra techniques. This paper will present two methods for optimizing indefinite functions with linear constraints, and computational results obtained using each method.
Descriptors : *OPTIMIZATION, *SPECIAL FUNCTIONS(MATHEMATICS), GLOBAL, COMPUTATIONS, STOCHASTIC PROCESSES, MATRICES(MATHEMATICS), ALGEBRA, PARALLEL PROCESSING, COSTS, SOLUTIONS(GENERAL), SEARCHING, SUPERCOMPUTERS, NONLINEAR PROGRAMMING.
Subject Categories : Numerical Mathematics
Operations Research
Distribution Statement : APPROVED FOR PUBLIC RELEASE