Large Scale Reasoning Using Allen's Interval Algebra

This paper proposes and evaluates a distributed, parallel approach for reasoning over large scale datasets using Allen's Interval Algebra (IA). We have developed and implemented algorithms that reason over IA networks using the Spark distributed processing framework. Experiments have been conducted by deploying the algorithms on computer clusters using synthetic datasets with various characteristics. We show that reasoning over datasets consisting of millions of interval relations is feasible and that our implementation scales effectively. The size of the IA networks we are able to reason over is far greater than those found in previously published works.