Traditional multi-criteria decision making (MCDM) algorithms are used in the handover of user equipment (UE) in an Ultra Dense Network (UDN). UDN refers to the increased density of the Radio Access Technologies (RATs) in a region which leads to the overlapping of the areas covered by individual RATs. MCDM algorithms such as TOPSIS, PROMETHEE and SAW are used to initiate handovers between these RATs based on the parameters obtained by the UE from each of the overlapping networks. However, initiating a handover abruptly and frequently, in case of availability of a new RAT without any thresholding technique proves to be unfriendly to the system resources. This can degrade the performance of the system. In this paper, a thresholding approach to the handover procedure is integrated to the MCDM process for the selection of RATs. First, an application-specific approach has been used in the selection of weights using the analytical hierarchy process which, is depending upon the application being used by the user. Then the ranking of the available RATs is done using the various MCDM algorithms and depending on the threshold specified for a handover, a decision is made whether to perform the handover process or not. In the case of streaming class of network traffic, the proposed method improves the performance of the system and reduces the handover by 13.14%, 19.35% and 8.62% of RAT modifications for TOPSIS, PROMETHEE and SAW respectively.

传统的多标准决策（MCDM）算法用于超密集网络（UDN）中的用户设备（UE）切换中。UDN是指某个区域中无线接入技术（RAT）密度的增加，导致各个RAT覆盖的区域重叠。诸如TOPSIS，PROMETHEE和SAW之类的MCDM算法用于基于UE从每个重叠网络获得的参数来启动这些RAT之间的切换。但是，在没有任何阈值技术的情况下，在新RAT可用的情况下，突然频繁地启动切换被证明对系统资源不友好。这会降低系统的性能。在本文中，切换过程的阈值方法已集成到用于选择RAT的MCDM过程中。第一，在使用权重的选择中，使用了一种分析层次过程，该过程取决于用户所使用的应用程序。然后，使用各种MCDM算法对可用RAT进行排名，并根据为切换指定的阈值，确定是否执行切换过程。对于网络流量的流式传输，所提出的方法提高了系统性能，并分别减少了TOPSIS，PROMETHEE和SAW RAT修改的切换的13.14％，19.35％和8.62％。然后，使用各种MCDM算法对可用RAT进行排名，并根据为切换指定的阈值，确定是否执行切换过程。对于网络流量的流式传输，所提出的方法提高了系统性能，并分别减少了TOPSIS，PROMETHEE和SAW RAT修改的切换的13.14％，19.35％和8.62％。然后，使用各种MCDM算法对可用RAT进行排名，并根据为切换指定的阈值，确定是否执行切换过程。对于网络流量的流传输类别，所提出的方法提高了系统性能，并分别减少了TOPSIS，PROMETHEE和SAW的RAT修改切换的13.14％，19.35％和8.62％。