Multi-access edge computing (MEC) has emerged as a promising technique for low-latency services, in light of its proximity to users and embedded cloud computing capability. In order to improve the network efficiency and fairness, it is crucial and challenging to jointly optimize resource management and user association mechanism with the consideration of heterogeneous services and network conditions. Moreover, the existing studies always neglect the heterogeneities of services on the requirements of both computation capability and storage capability. To solve this issue, we derive a strategy to improve the overall delay-aware performance of heterogeneous services with the MEC capability of computation and storage and the choices of the users' association. Accordingly, a coalition-game-based algorithm is proposed to form user coalitions for association scheme and resource sharing policy. In particular, we mathematically present that the proposed algorithm is capable of convergence and optimality. The simulation results also show that our algorithm gets a good performance efficiently. Furthermore, it reduces the weighted sum of delays of users by average 27.8% and 82.1%, while continuously improving delay-aware fairness, compared with those of the priority-based assignment scheme and the nearest assignment scheme, respectively. INDEX TERMS Multi-access edge computing, resource allocation, user association, coalition game.

There are frequent population flow and complex spatial structures in suburban villages. Understanding the spatial characteristics of population activities in suburban villages helps to coordinate the relationship between urban and rural areas and guide the development of suburban villages and the formulation of sound policies. Taking the rural area of Qin and Han New City as the research object, this paper constructs a population time-space analysis framework of “population attribute-activity characteristics-spatial analysis” based on cellphone signaling data. According to the characteristics of the population activity curve, K-means clustering algorithm was used to classify rural space and analyze their characteristics. This study has shown that migrants, who are showed as young and energetic, account for 49.8% of the local registered population per day. Bidirectional flow of residents and commuters is generally presented in urban and rural areas. The urban-rural relation curve was characterized by “double peaks”. The changes in the population in each village and the intensity of urban-rural relation were affected by location, industry and land use. The village population activity curve was classified into three categories, and nine characteristic villages are formed combined with the activity function. The research results can provide a scientific basis for urban and rural planning, spatial planning, industrial guidance and the facility layout.

No abstract

No abstract

No abstract