Core Concepts
The authors present a new general method called "piecewise analysis" to analyze the running time of branching algorithms. They apply this method to improve the running time analysis of two 17-year-old algorithms that solve the 4-Coloring and #3-Coloring problems.
Abstract
The paper introduces a new general method called "piecewise analysis" for analyzing the running time of branching algorithms. The key idea is to divide the instance space into groups based on a similarity ratio and then analyze each group separately. This allows the authors to take advantage of different intrinsic properties of instances with different similarity ratios, unlike the traditional measure & conquer (M&C) approach which uses a single measure for the entire instance space.
The authors showcase the potential of piecewise analysis by reanalyzing two 17-year-old algorithms from Fomin et al. (2007) that solve the 4-Coloring and #3-Coloring problems. The original analysis gave running times of O(1.7272^n) and O(1.6262^n) respectively, while the new piecewise analysis improves these to O(1.7215^n) and O(1.6232^n).
The key steps are:
Define a similarity ratio that divides the instances into groups.
Analyze the running time within each group separately using M&C.
Take the maximum running time across all groups as the final upper bound.
For 4-Coloring, the authors show that increasing the number of pieces further improves the running time, demonstrating the flexibility of piecewise analysis.