Accelerating the Requirement Space Exploration through Coarse-Grained Parallel Execution

Abstract

Abstract. The design and analysis of complex systems need to determine suitable configurations for meeting requirement constraints. The Monotonic Indices Space (MIS) method is a useful approach for monotonic requirement space exploration. However, the method is highly time and memory-Consuming. Aiming to the problem of low efficiency of sequential MIS method, this paper introduces a coarse-grained parallel execution mechanism to the MIS method for accelerating the process of requirement space exploration. The task pool model is used to receive and deploy hyperboxes for work balancing. To validate our approach, the speedup is estimated by a mathematical analysis and then an experiment is conducted in a PC cluster environment. The results show that high speedup and efficiency is achieved through our approach.