Research Article

State Aware-Based Prioritized Experience Replay for Handover Decision in 5G Ultradense Networks

Table 2

Comparisons of different handover methods according to their characteristics.

Ref.Problems and scenariosMethodContributionsSimulationsKPIs

[6]Coordinated multipoint handover in 5G UDNUser-centric CoMP handover schemesCharacterize the movement trend through dwell timeNumerical simulationHOR; Th
[7]Handover triggering policy in 5G UDNClustering-based RLMultiple decision criteria-based handover triggering mechanismMATLAB; SDHON; HOF; PPE; Th; latency
[12]Handover failures and ping-pong effect in HetNetHO triggers mechanismRecursive least squares-based SINR prediction methodSDHOF; PPE
[15]Handover decision issue in LTE-AAHP-TOPSIS; Q-learningUE rank; the optimal triggering points of HOMATLAB; SDHOR; PPE
[17]Handover in HetNetsMarkov-based handover strategyContext-aware handover policiesMonte Carlo simulationsCapacity
[22]Handover in 5GDNNReduce the handover problem to a classification problemSDRLF; PPE
[23]Frequent handover in 5G UDNFrequent handover mitigation algorithmDwell time estimation; user detectionNS3;HON; Th
[27]Handover decision in HetNetsQ-learningQ-learning-based handover decisionSDCost; utility
[30]Frequent handover in UDNDQNSDN-based UDN architecture; DQN-based handover decisionMininet;HOR; Th
[31]VHO in HetNetsES-DQNTraining the parameters of main Q-network with ESPython; SDHOF; Th; delay
[33]Frequent handover in 5GDouble DRLTrajectory-aware HO optimization approachWireless Insite software; SDHON; Th
[34]Handover decision in HetNetsDistributed DRLMDP formulation; distributed DRLSDHON; energy cost
[36]Handover in UDNA3CMobility pattern-based user clustering; A3C-based HO policySDHOR
[37]Handover and power allocation in HetNetsMultiagent DRLProximal policy optimization; cooperative multiagent DRLSDHOR; Th
[43]Load balancing and handover in 5G UDNLB-DDQNLoad balancing strategy; load coefficient; load balancing evaluation functionPython; SDHOR; Th
This paperHandover decision in 5G UDNSA-PERState aware; analysis of dwell time; the relationships between VHO and HHOPython; SDRLF; HON; PPE; Th

VHO: vertical handover; HHO: horizontal handover; HON: handover number; HOR: handover rate; Th: throughput; HOF: handover failure rate; RLF: radio link failure; PPE: ping-pong rate; SD: simulated data.