Calculating uncertainty intervals in approximate equation systems Original Research Article

This note suggests using an integer (binary) programming approach to the problem of finding intervals of uncertainty in systems of approximate equations. The current approaches—interval arithmetic, reformulating the problem as a simultaneous equations problem, and linear programming—become cumbersome and unwieldy for problems of even several variables, or when the feasible solutions are spread over several orthants in View the MathML sourcek. In contrast, the integer programming approach represents a generalization strategy to deal effectively with the combinatorial optimization nature of the problem. It can also be implemented with modest computational resources.