Comparison of Range Computation Methods
The problem of range computation is: Given an expression denoting a real-valued function f and intervals for all variables contained in the expression, compute an interval I such that f applied to elements in the intervals will always be in I and such that I is as small as possible. For example, for a function given by the expression xy and interval [0,1] for x and [-2,1] for y, a possible answer is the interval [-2, 1]. This problem can be solved using interval arithmetic, affine arithmetic, or linear relaxations. The goal of the thesis will be to implement and compare such methods.