Table 4.

Characteristics of selected resource allocation techniques in 5G.

Ref	Algorithm/Scheme/Strategy	Problem Addressed	Improvements/Achievements	Limitations/Weakness
[89]	Cooperative Online Learning Scheme	Extreme interference between the multi-tier users.	Maximizes spectral efficiency data rate by ensuring QoS.	Limited to tier and for downlink.
[90]	Game-theoretic approach	Cross-tier interference.	Improves spectral efficiency.	Limited to two tiers only.
			Energy efficiency.	Does not support uplink.
			Sum rate.
[91]	Genetic Algorithm Particle Swarm Optimization-Power Allocation (GAPSO-PA)	The allocation of power in heterogeneous ultra-dense networks.	Reduces the system outage probability.	Solves non-linear optimization.
[92]	Estimation of Goodput based Resource Allocation (EGP-BASED-RA)	Enhance Goodput (GP): (a specific metric of performance).	The performance of the UFMC system was boosted.	Limited to a particular packet format.
[93]	The social-aware resource allocation scheme	D2D multicast grouping;	Fairness.	Working on limited parameters.
		Ineffective D2D links.	Throughput.
			Has substantial benefits over other algorithms.
[94]	PGU-ADP algorithm	Dynamic virtual RA problem.	Drastically minimizes the outage probability.	Considers specific slice rate.
		Expansion of the total user rate.	Enhances user data rate	Slice state.
				Downlink only.
[95]	Efficient Resource Allocation Algorithm	Enhance system capacity and maximum computational complexity.	Improves system capacity.	Power allocation is done on the sub-carriers of the fixed group.
			Minimizes complexity performance.	Limited parameters.
[96]	GBD Based Resource Allocation Algorithm	Enhances allocating algorithm’s efficiency.	Total throughput achieved 19.17%.	Parameters are not suitable in all circumstances.
			Average computational time 51.5%.
			GBD with no relaxation by30.1%.
[97]	Multitier H-CRAN Architecture	Lacking intelligence perspective using existing C-RAN methods.	Manages spectral resources efficiently.	Necessary to improve bits of intelligence.
			Enhances control.
			End-to-end optimization.
			Ensures QoS by 15%.
[98]	Bankruptcy game-based algorithm	Resource allocation and inaccessibility of wireless slices.	Enhances resource utilization.	Focused on the cloud—RAN.
			Ensures the fairness of allocation.	Limited to specific parameters and slices.
[99]	BVRA-SCP Scheme	Enhancing service demands like low latency, enormous connection, and maximum data rate.	Beneficial resource utilization.	Limited to dynamic IoT-specific metrics.
			Low computational complexity.
			Supports dynamic IoT. slicing architecture.
			Improves efficiency and flexibility.
[100]	VNF-RACAG Scheme	Settlement of virtualized network functions (VNF).	The gain in end-to-end delay.	Limited parameters.
[101]	Hybrid DF-AF scheme	Promising to incorporate various wireless networks to deliver higher data rates.	Attains the concave envelope of the maximum between AF rate and DF rate.	Limited parameters are considered.
			Substantial gains for RFDRC.
[102]	Cooperative resource allocation and scheduling approach	Scheduling and resource allocation problems.	Decreases transmission collision probability.	Only for URLLC traffic.
			Enhances the reliability of upcoming 5G.	Considers limited parameters.
			Enhances vehicle-to-everything (ev2x) communications.
[103]	SWIPT framework	Low energy efficiency and high latency.	Maximizes energy efficiency.	Limited to downlink.
			Effective capacity.	Considers limited metrics.
[104]	The device-centric resource allocation scheme	Declining of network throughput and raises delay in resource allocation.	Reduces load at the BS up to 35%.	Improvement is required in intelligent resource allocation.
			Better performance.	Power efficiency was neglected.
[105]	Distributed Resource Allocation Algorithm	Resource allocation and interference management in 5G networks.	Efficient higher data rate results.	Limited to uplink only.
				Limited parameters were used.
[106]	Unified cross-layer framework	Physical layer modulation format and waveform, resource allocation, and downlink scheduling.	Enhances spectral efficiency using FBMC/OQAM.	Limited to specific parameters and frequency.
[107]	Dynamic joint resource allocation and relay selection scheme	Relay selection and downlink resource allocation.	Low computational complexity.	QoS neglected.
				Limited metrics are considered.
[108]	Low-Complexity Subgrouping scheme	Radio resource management of multicast transmissions.	Improves the Aggregate Data Rate (ADR).	Focused on data rate only.
			Ensures performance up to 9%.	QoS neglected.
				Limited parameters.
[109]	Joint Edge and Central Resource Slicer (JECRS) framework	Requires distinct resources from the lower tier and upper tier.	Satisfies latency and resource requirements.	Needs to support the NFVO.
			Guarantees communication and computing.
[110]	TCA algorithm	MTC devices are battery restricted and cannot afford much power consumption needed for spectrum usage.	Less complex.	N/A
			Achieves better performance.
[111]	IHM-VD algorithm	Power allocation and channel allocation issue.	Outperforms energy efficiency.	Focuses on specific parameters and particular domain.
			QoS requirements.
[112]	Centralized approximated online learning resource allocation scheme	The inter-tier interference among macro-BS and RRHS; and energy efficiency.	Ensures interference mitigation.	Limited to inter-tier interference mitigation.
			Maximizes energy efficiency.	Limited to specific parameters.
			Maintains QoS requirements for all users.
[113]	Spectrum resource and power allocation scheme	Emphasize on a fair distribution of resources in one cell.	Boosts system performance.	Limited to user interference in a single cell.
				Not suitable for multiple cell interference.
				QoS neglected.
[114]	Tri-stage fairness scheme	Resource allocation problem in UDN having caching and self-backhaul.	Improved flexible access and backhaul link resource allocation.	Particularly uses caching.
				Limited parameters are used.
				QoS is neglected.
[115]	Fronthaul-aware software-defined resource allocation mechanism	Overhead generated using a capacity-limited shared fronthaul.	Throughput enhancements.	Limited to in-band fronthaul.
			Delay reductions.	Limited parameters are used.
[116]	Heterogeneous statistical	Heterogeneity issues.	Efficient QoS across MIMO-OFDMA based CRNS.	Domain-specific.
	The QoS-driven resource allocation scheme			Limited parameters are used.
				Limited to effective capacity.
[117]	Nondominated sorting genetic algorithm II (NSGA-II)	Unable to get optimal results concurrently.	Performance.	Limited to ultra-dense network.
			Analyzes computational and convergence complexity.	Limited to downlink.
[118]	Joint access and fronthaul radio resource allocation	Downlink energy efficiency (EE) and millimeter-wave (MMW) links in access and fronthaul.	The system sum rate is enhanced up to 50%.	Limited parameters.
			Using PD-NOMA and comp the sum rate was enhanced up to 40%.	RAN-based only.
				Limited to downlink.
[119]	Double-sided auction-based distributed resource allocation (DSADRA) method	Intercell and inter tier interference.	User association satisfaction.	QoS not considered.
			Maximum output.	Limited for small cell only.
[120]	Joint power and reduced spectral leakage-based resource allocation	Interference from D2D pairs.	Reduces spectral leakage to nearby RBS.	QoS neglected.
			Ensures maximize signal-to-interference-and-noise ratio (SINR).	Limited parameters.
			Enhances overall throughput
[121]	Branch-and-bound scheme	Latency-optimal virtual resource allocation.	Enhances serviceability.	Limited to backhaul.
			Network load balance.	Limited parameters.
				Neglects energy efficiency.
[122]	The learning-based resource allocation scheme	To achieve high system capacity better performance in terms of effective system throughput.	More efficient in terms of system performance.	Limited to user’s position information.
[123]	Resource allocation method with minimum interference for two-hop D2D communications	Interference which reduces network throughput.	Enhances interference and throughput.	Limited parameters.
				Priority-based allocation block.
[124]	Multiband cooperative spectrum sensing and resource allocation framework	Energy consumption for spectrum sensing.	Satisfies the QoS requirement.	Channel fading changes over time.
				Mobile IoT nodes do not consider.
[125]	Channel-time allocation PSO Scheme	To acquire gigabit-per-second throughput and low delay for achieving and maintaining the QoS.	Encounters the growing requirements of applications.	Especially for multimedia traffic.
			Converged and high-capacity networks such as 5G.	Certain metrics are considered.
[126]	Heterogeneous (high density)/hierarchical (low density) virtualized software-defined cloud RAN (HVSD-CRAN).	Density of users.	Encounters variety of tradeoffs in resource management objectives such as cost, power, delay, and throughput.	Limited resource allocation in dense users.
				Certain parameters are used.
[127]	Mini slot-based slicing allocation problem (MISA-P) model	The probability of forming 5G slices.	Spectral efficiency and feasibility.	Limited parameter.
				Support single slot-based model.
				Limited for EMBB and URLLC traffic.
[128]	A joint resource allocation and modulation and coding schemes	Requirement of extremely low latency and ultra-reliable communication.	Achieves low error rates.	Only for URLLC traffic.
			Minimizes resource consumption.	Reserves resources for the first transmission.
[129]	QoS/QoE-aware relay allocation algorithm	Neglects temporal requirements for optimum performances.	Better performance for mean time to failure (MTTF).	Working based on different priorities.
			Average peak signal-to-noise ratio (PSNR).	Considers specific parameters.
			Average energy consumption.
[130]	The learning-based resource allocation scheme	Interference coordination complexity and significant channel state information (CSI) acquisition overhead.	Better effective system performance.	Accuracy varies as per user positions.
				Neglects throughput and QoS.
[131]	Device-to-device multicast (D2MD) scheme	Improving spectrum and energy efficiency and enabling traffic offloading from BSs to device.	Throughput enhanced.	Lack of attention to mobile users.
				Neglects selection of sharing mode and content caching in D2MD.
[132]	Constrained deferred acceptance (DA) algorithm and a coalition formation algorithm	The interference management among D2D and current users.	Enhances performance.	Limited coverage area.
			Throughput, fairness, and admitted users.	Neglects reliability and security.
[86]	Novel resource allocation schemes (hybrid resource management)	Energy efficiency and consumption.	QoS threshold and power budget are ensured.	Lack of attention to delay and overhead.
[133]	Orthogonal multiple access (OMA) and relay-assisted transmission schemes.	Jointly optimize the block length and power allocation for reducing error probability.	Improves performance.	Emphasis on short packet transmission only.
				QoS is neglected
[134]	Joint user association and Power Control algorithm	Optimizing power control and user association schemes.	Achieves higher energy efficiency performance.	Lack of attention to fairness and channel state information.
[135]	Successive convex approximation (SCA) based alternate search method (ASM)	Raise the total sum rate of users.	Enhances the performance of the system.	Lack of attention to fairness.
			Ensures the potential of SCMA.	Limited parameters are used.
[136]	An online learning algorithm for resource allocation	Inter-tier interference among RRHS and macro-BSs, and energy efficiency.	Enhances the energy efficiency.	Priority-based allocation of the resource block.
			Maintains users’ QoS.	Limited parameters are focused.
[137]	Joint resource block (RB) and power allocation scheme	Enhance fairness in data rate among end-users.	Low complexity.	Limited to femtocell only.
			Higher spectral efficiency.	Interference inside femtocell not considered.
[138]	Hybrid multi-carrier non-orthogonal multiple access (MC-NOMA)	Achieve the SE-EE tradeoff having minimum rate requirement of each user.	Outperforms both NOMA and OMA.	Decreases performance while adding more users.
			Enhances the tradeoff between system efficiency and user fairness.	Complexity.
[139]	Stackelberg game model	High inter-cell interference (ICI) and less energy efficiency.	Feasible and promising.	Focuses on limited parameters.
				Neglects intra-cell interference.
[140]	Virtual code resource allocation (VCRA) approach	Reducing the collision probability.	Reduces the collision probability.	Improves the code.
			Enhances efficiency.	Access to devices is according to priority.
[141]	Deep reinforcement learning -unicast-multicast resource allocation framework (DRL-UMRAF)	High-quality services and achieving green energy savings of base stations.	Improves energy efficiency.	Limited services framework.
			QoS requirements.	Limited to the number of cells and layers.
[142]	Deep reinforcement learning-based intelligent Up/Downlink resource allocation	The high dynamic network traffic and unpredicted link-state change.	Performance improvement.	Lack of attention to overhead.
			Packet loss rate and network throughput.	QoS neglected.
[143]	Joint computation offloading and resource allocation scheme	Complete network information and wireless channel state.	Outperforms energy consumption.	Limited to a specific parameter.
				QoS is neglected.
[144]	Deep neural network-Multi objective Sine Cosine algorithm (DNN-MOSCA)	Achieving better accuracy and reliability.	Better performance.	Spectral efficiency was neglected.
			Improves fairness, throughput, and energy efficiency.
[145]	The improved resource allocation algorithm	Improving QoS requirements in MTC.	Expressly improves the outage and success probability.	Prioritizes access for MTC devices.
				Limited parameters are considered.
[146]	Resource Allocation Algorithm	The interference to 5G cellular users (CUs) related to QoS.	Improves the cellular users’ channel capacity.	Limited parameters are considered.
			Guaranteeing QoS of the CUs.	Only for uplink.
[147]	Genetic algorithm- intelligent Latency-Aware Dynamic Resource Allocation Scheme (GI-LARE)	Efficient radio resource management.	GI-LARE outperforms these other schemes.	Divides traffic into 2 categories.
				Downlinks only.
				Specific parameters were used.
[148]	A Low-complexity centralized packet scheduling algorithm	Downlink centralized multi-cell scheduling.	Improves URLLC latency.	Neglects inter-cell interference.
			Achieves gains of 99% and 90% URLLC latency.	Considers only URLLC traffic.
[149]	Smart queue management method	QoS of end-to-end real-time traffic.	Confirms better end-to-end communication QoS of the real-time traffic.	Not for all IoT critical services.
			The average end-to-end communication delay was reduced.	Neglects other relevant parameters.
[150]	Proposed Optimal Resource Allocation Algorithm	The optimization problem in mixed-integer nonlinear programming (MINLP).	Improves throughput.	Wi-Fi or LTE only.
			Guarantees QoS of Wi-Fi user equipment.	Limited parameters are used.
				Good in one scenario only.
[151]	A novel packet delivery mechanism	Issues related to using CoMP for URLLC in C-RAN architecture.	Resource utilization.	Limited for URRLC traffic.
			UE satisfaction.	Lack of attention to overhead.
[152]	Distributed joint optimization algorithm for user association and power control	Improve total energy efficiency and reduce the inter-cell and intra-cell interference.	Effective and robust dynamic communication environment.	Limited to two-tier.
				Lack of attention to overhead.
[153]	Pollaczek–Khinchine formula based quadratic optimization (PFQO)	Inaccurate transmission recovery delay of URLLC multi-user services.	Bandwidth saving.	Lack of attention to retransmission timing.
			Packet length distributions.	Specific parameters.
[154]	An outer approximation algorithm (OAA)	Multiple interferences, imbalanced user traffic load.	Mitigating interference.	Lack of attention to QoS.
			Traffic offloading to address traffic imbalances.	Latency.
			Sum-rate maximization.
[155]	Joint Power and Subcarrier Allocation	URLLC reliability and network spectral efficiency.	Improves the spectral efficiency.	Limited to a single cell.
			URLLC reliability.	Not allocated slices in multiple cells.
				Neglects overhead.
[156]	Weighted Majority Cooperative Game Theory Based Clustering	Increase interference, improper utilization of resources.	Power consumption decreases up to 30%.	Fairness is not considered.
			SINR and spectral efficiency are increasedup to 40% and 45%, respectively.	Prioritizes small cells based on weight.
[157]	Bee-Ant-CRAN scheme	Design a logical joint mapping among RRHS and User Equipment (UE) and RRHS and BBUS too.	Improves the spectral efficiency as well as the throughput.	Neglects the effect of virtual BS.
				Lack of attention to energy efficiency.
[158]	Noncooperative game theory-based user-centric resource optimization scheme	Enhance the coverage probability and sum rate.	Improves the sum rate.	Limited to single macro cell scenario.
			Outage probability.	Neglects energy efficiency.