Worker Recruitment Mechanism Based On Spatio-Temporal Crowdsourcing In Edge Cloud Environment

In recent years,due to the rapid development of Internet of Things technology and the widespread application of various sensing devices,Spatio temporal Crowdsourcing(SCS)has developed into a new paradigm for obtaining sensing information and providing public services,used to recruit workers who can collectively perform large-scale tasks through intelligent devices.Worker recruitment is one of the core components of spatiotemporal crowdsourcing systems,and recruiting efficient workers often provides high-quality perceptual data,thereby improving service quality.One of the issues studied in this article is how to develop appropriate worker recruitment strategies and maximize task coverage under certain cost constraints.With the emergence of edge computing technology,edge computing sinks a large number of computing and storage resources to edge servers near the user end,which makes the requirements for the computing capacity and load capacity of edge servers become higher and higher.In practical crowdsourcing applications,workers often choose nearby edge servers to upload perception data.If a large number of workers unload computing tasks to the server near the edge server at the same time,it may lead to server overload or even paralysis.Therefore,how to select suitable workers to maximize task coverage under certain cost constraints,and how to balance the conflict between edge server resource consumption and service quality is another issue studied in this article.In the current stage of worker recruitment research,most researchers only consider a single situation online or offline and do not fully utilize task costs to recruit more efficient workers,resulting in resource waste.In addition,in terms of task offloading research on edge servers,existing algorithms cannot adapt well to complex crowdsourcing environments,and there may be conflicts between resource consumption and service quality.Therefore,this article has conducted research on existing problems and challenges,mainly including:1.In order to maximize the task coverage under the limited budget,a two-stage worker recruitment framework named BW Selector is proposed,which employs workers in two stages.In the offline phase,this paper proposes an opportunity aware worker recruitment algorithm.First,the Voronoi diagram is used to divide the task area,and then a prediction model based on long and short term memory(LSTM)is established to predict the movement trajectory of workers,which solves the cold start problem in the traditional SCS system.In the online phase,in order to recruit efficient workers,a participation aware employee recruitment algorithm based on adaptive threshold selection is proposed.2.To balance the conflict between resource consumption and service quality,through considering the learning generation unloading strategy among multiple Deep Neural Networks(DNN),a DTOO algorithm is proposed to obtain an approximately optimal task unloading strategy to resolve the conflict between resource consumption and quality of service,and a stack based unloading strategy is proposed.The resource ranking method allocates computing resources reasonably to reduce the probability of task failure.