An Efficient Task Synthesis Method Based on Subspace Differential Patterns for Arrangements of Event Intervals Mining in the Avionics Cloud System Architecture

Abstract

Avionics Cloud is a new multi-platform avionics system architecture that provides dynamic access, resource pooling, intelligent scheduling, on-demand service and other cloud computing features. Using Avionics Cloud to rationalize the order of multi-flight platform task execution and realize multitask synthesis is a challenging problem. In this paper, we propose an Efficient Task Synthesis Method based on Subspace Differential Patterns for Arrangements of Event Intervals Mining-DiMining. For tasks executed in a multi-platform Avionics Cloud system with dynamic characteristics of time intervals, DiMining is proposed. The algorithm mines the differential frequent task execution event interval patterns related to execution efficiency from the scenario dataset with high execution efficiency and the scenario dataset with low execution efficiency in order to identify key task patterns related to execution efficiency and improve the task synthesis design efficiency of the multi-platform Avionics Cloud system. Furthermore, in order to improve the mining efficiency of the algorithm, this algorithm designs a variety of pruning strategies to ensure that two differential time interval patterns with high and low functional execution efficiency are mined at one time without preserving the set of candidate items. The experimental results show that the DiMining algorithm is more efficient than the traditional algorithm on the open dataset. The DiMining algorithm is used to mine the 350-field high-efficiency operation scenario dataset and the 350-field efficiency operation scenario dataset under the constructed typical UAV cluster co-detection task scenarios. Based on the simulation results, the DiMining algorithm is able to effectively support the design of multi-platform Avionics Cloud system task synthesis architecture and improve the efficiency of UAV cluster collaborative detection.