Research On Multi-task Offloading Schemes In Mobile Edge Computing Networks

With the rapid development of mobile communication and internet technologies,more and more computing-intensive and delay-sensitive services appear in intelligent terminals,such as face recognition,augmented reality and virtual reality.These services have the characteristics of high computation complexity,energy consumption,and delay sensitivity,and they will also bring some great challenges to the computation power and standby time of mobile devices while increasing the network load.To tackle these challenges,ultra-dense networks equipped with mobile edge computing(MEC)are regarded as a promising option.In such networks,in order to save the computation resources and reduce the energy consumption,the mobile terminals(MTs)can offload computing-intensive tasks to the edge computing servers for processing.However,it will cause additional transmission time and result in higher network delay.To deal with the relationship between energy consumption and delay,the main research contents of this paper can be listed as follows.1)Ultra-dense mobile edge computing networks with multiple users(MTs)and multiple tasks are considered,and the equal frequency band division is integrated to eliminate the inter-layer and intra-layer interferences.Under such a network model,the one-step and two-step computation offloading schemes are considered,and the partial offloading modes are used to deal with the computation tasks.In the one-step computation offloading,MTs can be associated with the macro base stations(MBSs)and directly offload computation tasks to MBSs for processing.In the two-step computation offloading,any MT firstly offloads the computation tasks to one small base station(SBS)for processing,and this SBS can further offload partial task to MBSs for processing.To improve the performance of multi-step computation offloading,we try to optimize the frequency(band)division factor used to eliminate the inter-layer interference.2)In the above-mentioned model,the resource allocation is firstly performed according to the CPU occupation ratios of tasks,and then the MT association,power control,computation offloading,and frequency resource allocation are jointly optimized.For a multi-device multi-task system,this paper formulate two optimization problems to minimize the weighted sum of energy consumption and delay,and network-wide energy consumption.3)The above-mentioned formulation problems are in a nonlinear and mixed-integer form,it is often difficult for us to find the closed-form optimum solutions of them.To solve the problem with minimizing the sum of weighted delay and energy consumption,this paper take account of adaptive particle swarm optimization(APSO).To this end,we must reasonably complete the particle coding before algorithm design.In order to verify the effectiveness of APSO algorithm,other algorithms are introduced for comparison in the simulation.4)To solve the problem with minimizing network-wide energy consumption,we develop an improved hierarchical adaptive search(IHAS)algorithm.To this end,this paper need to make appropriate changes to the gene coding,selection,crossover,and mutation according to the practical form of formulated problem.In addition,in order to highlight the effectiveness of designed algorithm,other existing algorithms are introduced for comparison in the simulation.