Task Allocation Based On Whale Optimization Algorithm In Crowdsensing Systems

With the increasing popularity of mobile terminals with various built-in sensors(such as cameras,accelerometers,microphones,gyroscopes,GPS,etc.),a new type of sensing mode called mobile crowdsensing has emerged.Unlike traditional static sensor networks,which arrange a large number of static sensors to collect data,mobile crowdsensing collects data through mobile terminals carried by crowds and can collect a large range of data at lower costs.Thus,mobile crowdsensing has been widely used in many fields such as transportation services,environmental services and so on.Nevertheless,it should be pointed out that how to efficiently assign tasks to workers is very important in mobile crowdsensing systems.Consequently,these task allocation problems have received widespread attention.By far,some researchers have studied the task allocation problem in crowdsensing systems based on the workers' attributes or tasks' attributes,but the attributes considered in these works are relatively simple and cannot fully reflect actual situation.In order to avoid the limitations of existing works,this dissertation deals with more complicated situations,and studies task allocation problems from the perspectives of workers' attributes,tasks' attributes as well as matching degree of workers' and tasks' attributes.Furthermore,the corresponding task allocation algorithms are proposed with the objectives of maximizing efficiency,maximizing revenue,and maximizing amount of completed tasks,respectively.The main works in this dissertation are summarized as follows:(1)The task allocation problem of maximizing efficiency based on workers' attributes.First,we consider the situation of workers with the elastic online time,which is described by the fuzzy opportunity constraint programming method.Then,the idle cost and delay cost are introduced according to delay or not,and the corresponding task allocation problem is given.Finally,based on the discrete whale optimization algorithm,the two-stage allocation algorithm is designed,and the performance of the designed algorithm is verified by simulation experiments.(2)The task allocation problem of maximizing revenue based on tasks' attributes.First,we consider the tasks with attributes such as sensing time,sensing positions,sensing contents,etc.Then,we analyze the coincidences of sensing time,sensing positions and sensing contents and define the spatiotemporal correlation and content correlation between tasks,respectively.Moreover,the periodic allocation framework is proposed based on the sensing periods which determined by multiple sensing periods of tasks.In the framework,the improved whale optimization algorithm is used to solve the corresponding task allocation problem,and the collaborative model is implemented to cope with the long-distance sensing tasks.Finally,the simulation experiments are conducted to assess the performance of our proposed allocation framework.(3)The task allocation problem of maximizing amount of completed tasks based on matching degrees between workers' attributes and tasks' attributes.First,the spatial matching degree and the skill matching degree between workers and tasks are defined based on position attributes and skill attributes,respectively.Then,considering the increasing experiences of workers brought by their unique learning ability,the change in workers' skills with the increasing experiences is defined,and the corresponding task allocation problem is given.Finally,we improve the whale optimization algorithm by the differential evolution algorithm,and the worker selection strategy is developed based on the improved whale optimization.Simulation experiments verify the performance of the designed strategyFocusing on the task allocation problems in crowdsensing systems,this dissertation carries out research from the perspectives of workers' attributes,tasks' attributes,and matching degrees between workers' attributes and tasks' attributes.Furthermore,the improved whale optimization algorithms are adopted to design the corresponding task allocation algorithms,and the relevant works can provide an important reference for the realization of crowdsensing systems.