Some important problems in the computing sciences or in engineering require that a certain performance metric or function be optimized while under conflicting constraints. An practical example of an optimization problem would be the design of an automobile engine to deliver maximum possible fuel efficiency while staying within size, weight, and cost constraints. Optimization problems are analyzed and solved by stripping away the inessential details of the problem domain to obtain an abstract specification of the problem. This requires a clear understanding of just what features of the problem domain are relevant and what are not--that is part of the challenge of meaningful abstraction.

The oldest and best-known technique used in optimization is linear programming. The word "programming" is in some respects an unfortunate choice here, since it does not connote programming in the sense of creating software programs or applications. It really means "planning," and...