Results
For the test cases that have been tried the solutions produced by our decomposition algorithm are close to the optimal solution, but the latter is not always reached. The reason behind this is that the decomposition algorithm is a sequence of iterations that may accumulate errors at different levels---lowering the tolerances used as stopping criterion in the submodels most of the time does not improve the solution. However, the configuration of the decomposition algorithm itself shows some impact on the solution: in phases 1 and 2 one may choose, for instance, to stop once the first submodel returns no improvement or continue until no more proposals are generated. Generating more proposals sometimes helps finding a better solution, but it also increases the number of times (sub)problems are solved and hence prolongates the solving time.
© 2001-2020 Fair Isaac Corporation. All rights reserved. This documentation is the property of Fair Isaac Corporation (“FICO”). Receipt or possession of this documentation does not convey rights to disclose, reproduce, make derivative works, use, or allow others to use it except solely for internal evaluation purposes to determine whether to purchase a license to the software described in this documentation, or as otherwise set forth in a written software license agreement between you and FICO (or a FICO affiliate). Use of this documentation and the software described in it must conform strictly to the foregoing permitted uses, and no other use is permitted.