Global SPACING Constraint (Technical Report)

TL;DR

This paper introduces a new global SPACING constraint for modeling temporally distributed events, providing theoretical insights, efficient algorithms, and demonstrating superior performance in a music composition application.

Contribution

It presents the first formal study of the global SPACING constraint, including theoretical properties, tractable cases, and efficient filtering algorithms, with experimental validation.

Findings

Filtering algorithms outperform existing methods in music composition tasks

Theoretical properties and tractable cases are identified for the new constraint

Efficient domain consistency filtering algorithms are proposed

Abstract

We propose a new global SPACING constraint that is useful in modeling events that are distributed over time, like learning units scheduled over a study program or repeated patterns in music compositions. First, we investigate theoretical properties of the constraint and identify tractable special cases. We propose efficient DC filtering algorithms for these cases. Then, we experimentally evaluate performance of the proposed algorithms on a music composition problem and demonstrate that our filtering algorithms outperform the state-of-the-art approach for solving this problem.