Abstract: In parallel distributed data processing frameworks like Spark and Flink, task scheduling has a great impact on cluster performance. Though task Scheduling has proven to be an NP-complete ...