Unformatted text preview: optimal solution to (8) would be to carry out a complete enumeration of the feasible set S in an intelligent way and pick the optimal feasible solution. Branch & bound algorithms consist of two main components: • • A successful partitioning of the feasible set into mutually disjoint subsets Si (branching) An algorithm is available to calculate (lower) bounds zi of the objective function on these subsets Si (bounding). Now, if the lower bound zi on the subset Si is higher than the optimal feasible solution found so far, we can completely rule out the optimal solution being in Si. In the worst case, however, we still have to enumerate all the solutions. We present the discrete tolerance allocation problem formulation in the next section and propose an optimal method for how to solve it in an efficient way. An Efficient Solution
4. THE DISCRETE TOLERANCE ALLOCATION PROBLEM 121 4.1. Problem formulation We now present our formulation of the discrete tolerance allocation problem. For each dimension i,...
View Full Document
- Spring '10
- The Land, ........., Computer-aided design, Joseph K. Davidson, M. GIORDANO