Improvement in the Quality of Solutions of a Heuristic Linear Decomposer for Index Generation Functions

This paper presents a method to improve a heuristic linear decomposer for index generation functions. Besides being fast and scalable, it also improves the solutions. This is done through the use of an efficient evaluation method that can find better solutions and can predict the quality of the solution. Experimental results show that the improved heuristic quickly finds exact optimum solutions that are not found by existing heuristics.