Power Control Techniques for Interference Management—A Systematic Review

Abstract

Applying optimal power control techniques in wireless networks is invaluable to mitigating interference among mobile devices. This review seeks to evaluate the depth and extent of the application of power control in 5G wireless networks through a systematic literature review. This review includes journal articles from 2018 to 2023 indexed in the Scopus, Web of Science (WoS), and IEEE Xplore databases. We used the following search string to search articles from each database: (*power control” AND “resource management” OR interference management AND 5G mobile communication). We obtained 3561 articles from the Scopus, WoS, and IEE Xplore databases with respective counts of 254, 728, and 2579. We paid attention to journal articles to ensure the quality of the review. After carefully assessing each record, we selected 770 journal articles using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. Using R package software, we performed analyses based on the number of citations, keyword co-occurrence, and trending topics. This review reveals that various power control taxonomies address interference in 5G wireless networks. The results confirm continuous growth in the study, signifying the need for further exploration.

1. Introduction

One of the primary design challenges in modern wireless communications is transmit power control. Power control involves using techniques and algorithms to manage and adjust transmit power from base stations and user equipment (UE). One possible scenario is that a high level of it may result in a high signal-to-interference-noise ratio (SINR), which may cause interference. Another likely scenario is that a lower SINR results in a poorly received signal [1,2]. We employ power control techniques to adjust the SINR dynamically. The benefits include extending the UE’s battery life and maintaining an acceptable quality of service (QoS) [3,4]. Quality of service, among other factors, is limited by interference [5]. The successful implementation of power control techniques offers the benefit of reducing interference among UEs [6]. In addition, other advantages include reducing co-channel interference [7], maximizing cell capacity [8], and minimizing UE mean transmit power [9,10,11].

1.1. Existing 5G Reviews on Power Control Techniques

Numerous review articles have been published to evaluate power control techniques in wireless networks. Table 1 shows review papers focused on power control techniques. We tabulated their focus and contributions to identify their relevance to the topic under review.

Table 1. Review papers on power control techniques.

Source	Year	Survey/Review Focus	Contributions
Siddiqui et al. [12]	2021	This paper comprehensively surveys interference issues studied in technologies like the Internet of Things, heterogeneous networks, massive multiple-input multiple-output (MIMO), and Relay Nodes.	The paper provides insights into interference management for next-generation networks. Furthermore, the review explores types of interference and techniques to minimize interference.
Banafaa et al. [13]	2024	The paper presents various challenges to integrating unmanned aerial vehicles (UAVs) in 5G and beyond networks. Challenges include the effective management of spectrum resources to address interference.	This survey outlines key challenges and research directions essential for realizing the full potential of UAV technology. Furthermore, it offers a foundation for decision making, collaboration, and driving advancements in UAV communication technologies.
Chataut et al. [14]	2020	The paper presents challenges concerning pilot contamination, channel estimation, precoding, user scheduling, energy efficiency, and signal detection in massive MIMO.	Discusses open research issues to direct future research in massive MIMO systems for 5G and beyond networks.
Merin et al. [15]	2023	The paper reviews non-orthogonal multiple-access (NOMA) technologies in 5G and beyond networks, including 6G networks. The review further discusses how interference is minimized in uplink and downlink NOMA systems.	The review discusses the barriers identified while implementing NOMA technologies in 5G and beyond networks. Future research possibilities are discussed.
Zhang et al. [16]	2018	The article provides a survey of prevalent IoT technologies deployed within licenced cellular spectrum and unlicensed spectrum. Furthermore, the article focuses on spectrum-sharing solutions, including shared spectrum, interference models, and interference management.	Identifies both the advantages and disadvantages of different IoT technologies. Furthermore, the article identifies challenges for the future IoT and suggests potential research directions.
Sufyan et al. [17]	2023	This presents a comprehensive survey of studies on the challenges of the evolution and implementation of 5G wireless networks that have been described in terms of enhanced data, channel capacity, and spectral efficiency.	This article reviews key emerging wireless technologies to increase channel capacity and spectral efficiency. This, in turn, leads to improvement in network performance and meets ever-increasing user demands.
Nwankwo et al. [18]	2018	This paper presents a comprehensive survey of the literature on self-interference management schemes required to achieve single-frequency full-duplex (FD) communication in wireless networks.	The survey presents the strengths and limitations of SIMS, the amount of SIC required for different access schemes, and key research areas for future work.
Douros and Polyzos [19]	2011	A systematic study on power control issues in wireless networks was presented in the work published from 1992 to 2011.	Pointed out some critical research directions that could be further pursued.
Budhiraja et al. [20]	2021	This study presents an in-depth survey of state-of-the-art non-orthogonal multiple-access (NOMA) variants with power and code domains as the backbone for interference mitigation, resource allocations, and quality of service management in the 5G environment.	Provides in-depth analysis of orthogonal multiple-access (OMA)- and NOMA-based applications using a systematic review methodology. A comparative analysis was carried out to provide further insights.

1.2. Current Review Limitations

The existing reviews have limitations, as illustrated in Section 1.1. Recent trends show an exponential increase in published articles, including review articles, in general [20]. Literature reviews are broadly classified as either “narrative” or “systematic” [21,22]. Literature reviews on power control have presented constitute narrative reviews. The authors could not identify reviews that addressed interference management. Furthermore, we identified two review papers: Douros and Polyzos [19] and Budhiraja et al. [20]. The authors in [19] presented a systematic review methodology for evaluating power control issues in wireless networks of the work published from 1992 to 2011. Reference [20] used a systematic review to compare the performance of OMA and NOMA-based applications in 5G networks. The authors could not identify related reviews on the current technologies, such as 5G wireless networks. Because of this limitation, the authors could not identify current research publications to address interference in 5G wireless networks. None mapped out opportunities or directions for future research on power control techniques in 5G wireless networks.

1.3. Research Objectives

The main aim of this review paper is to evaluate the current state of power control techniques in 5G wireless networks to provide a basis for new research topics. This aim is achieved by accomplishing the following research objectives:

• Identify the current publication trends in power control techniques to address interference in 5G wireless networks.
• Provide the state of the most influential authors and journal articles on power control techniques.
• Provide the current state of co-word occurrence in power control to address interference in 5G wireless networks.
• Point out power control areas in 5G wireless networks that need additional study and the mapping of opportunities and directions for future research.

1.4. Significance of the Study

A systematic literature review of the existing literature was conducted to present a comprehensive overview of the current literature on power control techniques on 5G wireless networks. This review can open new approaches to power control techniques to address interference. Implementing this successfully could bring efficient technique(s) to address interference. This review delivers a clear and comprehensive overview of available evidence on the topic under review.

1.5. Review Outline

The remaining sections of the paper are organized as follows. Section 2 discusses the systematic literature review methodology. The results obtained from Section 2 are presented in Section 3 (Analysis and Findings) to address the research objectives outlined in Section 1.3. This review is strengthened by discussing existing power control theories in Section 4. Section 5 presents the discussion and recommendations for future study. Section 6 presents the limitations of this study. Finally, the review is concluded in Section 7.

2. Systematic Review Method

In this study, the systematic literature review shown in [23,24] was used. The authors in [23,24] expound on the stages of conducting a successful literature review. The stages are (1) planning the review, (2) the identification of relevant literature, (3) data selection and data extraction, (4) data analysis and synthesis, and (5) reporting the review. The planning stage involves identifying the necessity of the review, stipulating research questions, and developing a review procedure. The conducting stage consists of the identification and selection of primary studies, as well as the extraction, analysis, and synthesis of data. The research findings should then be disseminated by reporting the review results. These steps are summarized in Figure 1. The following sections and subsections implement these steps for the systematic literature review.

Figure 1. Systematic literature review process. Adapted from [25].

2.1. Planning the Review

2.1.1. Contextualization of the Review

Contextualization presents an approach to research work. Furthermore, it gives credibility and motivation to conduct research [26]. In this study, we contextualized our study concerning prior reviews on power control studies to provide credibility and support to our review. Section 1 (Introduction) presented a thorough discussion of this stage. The following subsection will highlight the development of the review procedure.

2.1.2. Development of the Review Procedure

The review procedure is carried out in the same way that research design is carried out in social science studies [25]. The authors in [24] emphasized the need for a review procedure for rigorous systematic reviews. The authors in [27] endorse the need for a review procedure. Gomersall et al. [28] suggested the elements of the review procedure, including “the purpose of the study, research questions, inclusion criteria, search strategies, quality assessment criteria and screening procedures, strategies for data extraction, synthesis, and reporting”. The following paragraphs discuss how each procedure is applied in the context of this research.

(i)

The purpose of the review

This article investigates models that have been proposed in the evolving 5G networks to address resource management. The objective is to search for models in the literature developed to address resource management in 5G networks. The successful completion of this research will inform future research.

(ii)

Research questions

Based on the research objectives outlined in the Introduction section, this study examined the following research questions (RQs):

• RQ1: What are the current publication trends in power control techniques for addressing interference in 5G wireless networks?
• RQ2: Which are interference management’s most influential authors, countries, and journal articles?
• RQ3: What is the current state of collaboration involving power control to address interference in 5G wireless networks?
• RQ4: Which areas involving power control in 5G wireless networks need additional study? (This will point out opportunities and directions for future research.)

This study was carried out to answer these research questions by implementing the steps shown in Figure 1.

2.2. Identification of the Relevant Literature

2.2.1. Keywords and Database Search Filters

We developed a search strategy for this systematic search to identify the relevant literature. This search strategy targeted journal articles published in English from 2018 to 2023. The search strings from each database, including the keywords and the operators, were as follows: “power control OR power allocation AND resource management OR interference management AND 5g wireless network* OR fifth-generation wireless network*)”.

After running a search, we applied various filter options available for each database to refine and focus the results yielded. Each database has limits or filters, including publication years, document types, and content types. More detailed filters can be applied using the other options available. Table 2 presents the filters applied for each database to reach the refined article count for this review.

Table 2. Database search filters.

Refining Filters
Scopus	Web of Science	IEEE Xplore
Open Access: All Open	Quick Filter: Open Access	Quick Filter: Open Access
Publication years: 2018–2023	Publication years: 2018–2023	Publication years: 2018–2023
Document Type: Article	Document Types: Article
Source Type: Journal	Web of Science Categories: Telecommunication	Source Type: Journal
Keywords: Power Control; Resource Management; Optimization; and Mobile Telecommunication Systems	Citation Topics Meso: Wireless Technology; and Telecommunication	Topics: 5G Mobile Communication; Wireless Channels; Telecommunication power management

Scopus, WoS, and IEEE Xplore allow users to download the search results in different formats, including CSV and Excel (tab-delimited file). For each literature source, we determined its relevance using its title. From the title, if the content seemed to discuss power control modelling or techniques in 5G wireless networks, we obtained its full reference (including author(s), year, title, etc.), interference management approach(es) and technique(s), and the country. A summary of the observations from each article was made.

2.2.2. Sources for Literature Search

The search process entails a manual search of journal papers and conference proceedings from 2016 to 2023 to build our review of the recent literature. The authors in [29] emphasize the need to use more than one database for literature searches to maximize the likelihood of finding potentially relevant studies. However, no definite number of databases are recommended to provide a complete, objective, and unbiased relevant selection. This is due to the differences in coverage and indexing methods used [29]. Finally, we may use multiple databases for a literature review, not just one. This is because no single database can cover all the relevant and reliable sources for a topic and discipline. To broaden our scope, avoid bias, and enhance our research credibility, we used three databases in our search. Additionally, this enabled us to crosscheck and compare the results from the three databases to identify the most influential sources for our literature review. Suffice it to mention that there is no definite number of databases to use for literature searches. In this research, we limited the search to three databases as too many can make our search too complex to analyze. In this study, three databases, shown in Table 3, were used to search for the literature. The databases were selected because of their usefulness across various disciplines, including electrical engineering.

Table 3. List of databases.

Databases	Public Target	Link
Scopus	Scopus is Elsevier’s abstract and citation database of peer-reviewed journals in fields such as life sciences, social sciences, physical sciences, and health sciences	https://www.scopus.com/ (accessed on 28 October 2023)
Web of Science	Web of Science provides access to the most reliable, integrated, multidisciplinary research connected through linked content citation metrics from multiple sources within a single interface.	https://webofknowledge.com/ (accessed on 28 October 2023)
IEEE Xplore	IEEE Xplore provides web access to over three million full-text documents from some of the world’s most highly cited electrical engineering, computer science, and electronics publications.	https://ieeexplore.ieee.org/ (accessed on 28 October 2023)

2.3. Selection and Data Extraction

2.3.1. Literature Search Results

(i)

Scopus database search results

We first searched the Scopus database in August 2023. This search yielded a total of 254 articles. The retrieved articles were exported into Mendeley, and duplicates were removed. Articles were screened, as discussed, to include relevant articles and exclude nonrelated articles. Then, the full texts were investigated to identify relevant articles. All journal articles were read in detail. A total of 54 articles were eventually included in the review.

(ii)

WoS database search results

A search on WoS yielded a total of 728 studies. The retrieved articles were exported into Mendeley, and duplicates were removed. Articles were screened, as discussed, to include relevant articles and exclude nonrelated articles. Then, the full texts were investigated to identify relevant articles. All journal articles were read in detail. A total of 324 articles were eventually included in the review.

(iii)

IEEE Xplore database search results

A search on IEEE Xplore yielded a total of 2579 studies. The retrieved articles were exported into Mendeley, and duplicates were removed. Articles were screened, as discussed, to include relevant articles and exclude nonrelated articles. Then, the full texts were investigated to identify relevant articles. All journal articles were read in detail. A total of 392 articles were eventually included in the review.

Although not exhaustive, this profile produced many documents for further analysis. We identified 2029 potential studies from all databases before searching for duplicates and excluding them. After removing duplicates and screening, 770 journal articles were identified. A search on Scopus, WoS, and IEEE Xplore returned 54, 324, and 392, respectively. Thus, the search yielded a total of 770 journal articles. The search results are presented in Figure 2. This method is known as “Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)” [30].

Figure 2. PRISMA flow diagram procedure. Adapted from [31].

2.3.2. Inclusion and Exclusion Criteria

The procedure of a systematic literature review must be detailed in sufficient detail for literature reviews to be dependable and independently repeatable [27]. Other researchers can use this information to follow the same procedures and obtain the same conclusions. The report should describe the reasoning behind each inclusion and exclusion criteria [32,33], focusing on the inclusion and exclusion criteria. The following inclusion and exclusion criteria were defined:

(i)

The criteria for inclusion were the following:

• If the content of the article seemed to discuss modelling in 5G wireless networks, we obtained its full reference (including author(s), year, title, etc.), interference management approach(es) or technique(s), and the country.
• Articles were limited to those that included specific strategies for resource management in 5G wireless networks.
• Journal articles that mentioned power control in 5G wireless networks.
• This paper suggests collecting data from the following databases: Scopus, Web of Science, and IEEE Xplore Digital Library.
• Journal articles written in English.

(ii)

The criteria for exclusion are:

• Journal articles that described no resource management modelling or techniques in 5G wireless networks were excluded.
• Any study that is not about 5G wireless networks.
• The article does not describe any resource management models in 5G wireless networks.
• Studies before 2018.
• Published data not available.

After applying the inclusion and exclusion criteria, 2029 journal articles remained for the next step of the procedure.

2.3.3. Quality Assessment

Quality assessment (QA) is a stage that follows the screening for the inclusion stage. In this stage, researchers obtain full texts of the studies under review. The QA acts as a final stage to ensure that the full-text articles from each study are ready for data extraction and synthesis. In this study, we are not analyzing studies; our review is on power control in 5G networks. Thus, QA is only limited to the articles we obtained from the databases [25]. The authors read through the full text to carefully examine each journal article against the quality criteria (inclusion/exclusion). The full-text review also provided an opportunity for a final check on inclusion/exclusion. Studies that did not satisfy the inclusion criteria previously specified were excluded from the final literature list [25]: “the list of excluded papers should be retained for record-keeping, reproducibility, and crosschecking”. Finally, 770 journal articles related to the defined research questions were considered.

2.3.4. Data Extraction

One information specialist from the University of South Africa library and the two Professors playing supervisory roles in this study conducted initial data extraction. Each of the parties was involved independently. The lead author (AL) reviewed all final studies and iteratively generated the list of the literature. The characteristics of the studies collected included authors, year of publication, objective, study design, setting, intervention studied (if any), participant sampling methods, inclusion and exclusion criteria, data collection and analysis methods, participants’ characteristics, main results, and authors’ conclusion(s).

The data extracted from each study were the following:

• Publication source (author(s)).
• Year of publication.
• Power control taxonomy.
• Specific power control technique.
• Contributions.
• Limitations/future work.

2.4. Data Analysis and Synthesis

The analysis was carried out using Bibliometrix. Bibliometrix is the most popular R package (version R-4.4.2) and is used in many publications. It allows R users to import a bibliography database from SCOPUS or the Web of Science, stored either as a Bibtex (.bib) or Plain Text (.txt) file.

The network units of analysis for consideration to address the research questions (RQs) were bibliometric coupling (RQ1), citation analysis (RQ2), co-authorship (RQ3), and co-word analysis (RQ4), as shown in Table 4.

Table 4. Bibliometric units of analysis to address the research questions.

Research Question (RQ)	Units of Analysis	Aim of the Analysis
RQ1: What are the current publication trends in power control techniques to address interference in 5G wireless networks?	Data overview Main information Annual scientific production	To identify the main themes to quantify and visualize the thematic evolution of power control in 5G networks. The analysis focuses on citing documents [34] In addition, it helps us to understand the current state of the topic under study and to predict the direction of future research [35].
RQ2: Which are the most influential authors and articles on power control to address interference in 5G wireless networks?	Sources Most locally cited authors Most relevant sources Documents Most globally cited documents	To provide the relationships between authors and the leading publications. It tracks authors’ words, the impact of papers, and the trajectory of research ideas by measuring citation counts in critical research databases and online sources [36].
RQ3: Which authors’ keywords are leading the list in the study?	Documents Words (authors’ keywords) Most frequent words	To provide relationships among authors, countries, and research institutions. It examines interactions among scholars, presents intellectual collaboration among them, and informs the development of the research field [37].
RQ4: Which are the trending topics on power control to address interference in 5G wireless networks?	Documents Words (authors’ keywords)∙ Trend topics	To investigate the recurring themes in a research field and to identify possible future research directions. It gives interpretations of co-citation analysis (past) or bibliometric coupling (present) and predicts forthcoming trajectories (future) [8].

As pointed out in [34], bibliometric coupling is a relationship measurement that occurs when authors cite the same author. When authors share a high number of citations, the bibliometric coupling between them is more substantial. According to Aria and Cuccurullo [38], in bibliographic coupling, a relationship is established by authors of the relevant publications instead of a co-citation analysis, where the relationship is amongst authors who cite the documents under consideration. The documents under investigation serve as the primary source for analysis [34]. The benefit of the analysis is twofold. Firstly, the critical themes extracted from the documents will help quantify and depict the thematic development of power control schemes in 5G networks. Secondly, it aids in predicting the course of future research and the state of the art of the subject under investigation [35].

The objective way to measure the quality of a research paper is through co-citation analysis. This measures the impact the research makes in a particular field or area. This translates into the interpretation of the following units of analysis, including journals, authors, documents, research institutions, and countries. The relationships among and leading publications are identified by undertaking the analyses. Furthermore, the analysis aids in identifying the trajectory of research ideas [39]. In this study, the trajectories concern the application of power control schemes in 5G networks.

Co-authorship analysis assesses the relationships among authors and countries, their affiliations and other comparable effects on the advancement of the research area [40]. The analysis performed presents the following benefits. Firstly, scholars interact amongst themselves and, as a result, affect intellectual collaboration. This helps to advance knowledge in the area under study. Secondly, it informs the future development of the research field [41].

Co-word analysis is a method that looks at the actual content of the publication itself, using “keywords” as the unit of measure, as opposed to citation analysis, co-citation analysis, and bibliographic coupling, which use cited or cited publications as the point of analysis [42]. The keywords are frequently generated from “author keywords.” The examples shown by authors in [35] and [43] illustrate the application of co-word analysis. Like co-citation analysis, co-word analysis assumes that words frequently appear together and have a similar thematic relationship [44]. This analysis investigates the recurring themes in a research field and identifies possible future research directions. Furthermore, it gives interpretations of co-citation analysis (past) or bibliometric coupling (present) and predicts forthcoming trajectories (future) [45].

3. Analysis and Findings

This section presents the literature search’s findings on power control techniques based on the results obtained in Section 2. A bibliometric analysis of the literature obtained from the databases was conducted using the R package Bibliometrix (version R-4.4.2) [8] software. In this study, we considered four types of analyses, as discussed in Section 2.

3.1. Dataset

3.1.1. Main Information

After merging and removing duplicates from the Scopus, Web of Science, and IEEE Xplore databases, we obtained 54, 324, and 392 journal articles, respectively. These documents were collected from 85 (combined) sources published between 2018 and 2023 (until October 2023). Biblioshiny captured a total number of 2734 authors. Table 5 presents additional statistics of the key data used for bibliometric analysis.

Table 5. Overview of the main information about the data.

Description	Scopus	Web of Science	IEEE Xplore
Timespan	2018–2023 (until October 2023)
Documents	54	324	392
Sources (journal articles)	27	31	27
Keywords plus (ID)	499	505	381
Author’s keywords (DE)	181	1305	1302
Average citations per document	14.89	17.32	14.79
Annual growth rate %	−2.33	16.05	−6.32
Authors	200	1195	1339
Authors of single-authored documents	0	3	6
Single-authored documents	0	5	9
Co-authors per document	3.89	4.44	4.32
International co-authorship %	40.74	47.22	0

3.1.2. Annual Scientific Production

An accurate indicator of the current publication trend in a particular field of study is the number of articles published yearly. The trend analysis of the number of publications provides information on future research trends. Table 6 displays the distribution of publications that received power control over time and the total number of publications.

Table 6. Annual scientific publication for each database.

Year	Scopus	Web of Science	IEEE Xplore
2018	22	19	61
2019	15	29	95
2020	12	83	84
2021	9	94	68
2022	14	59	40
2023	17	40	44

For ease of analysis, we present Figure 3 to illustrate the annual scientific production of the journal articles. We observe similar patterns from the WoS and IEE Xplore databases. The WoS database follows the IEEE Xplore trend; however, its response has been delayed by a few months. Considering this, we may predict an increase in the trend of WoS beyond 2023. In comparison with other databases, Scopus indicates a slight increase in the number of articles. Several factors can cause these publication trends. This could result from changes in peer review processes or an increase or decrease in open access journals. Furthermore, there is a shift in funding when some journals receive more funding than others. A high impact factor and the number of citations could also be reasons for database publication trends.

Figure 3. Annual scientific production per database.

3.2. Sources

3.2.1. Most Relevant Sources

Table 7 shows the journals that published the most articles regarding power control from 2018 to 2023. The journal with the most papers published was “IEEE access” in all databases. The highest record publications from Scopus, WoS, and IEEE Xplore were 33, 147, and 330, respectively. The second most popular sources were “IEEE transactions on communications”, “IEEE transactions on wireless communications”, and “IEEE open journal of the communications society” with records of 8, 44, and 17, respectively. Clearly, “IEEE access” is ahead of each of them by a large margin, which indicates its high impact on power control techniques.

Table 7. Most relevant sources from Scopus, WoS, and IEEE Xplore.

Scopus		Web of Science		IEEE Xplore
Sources	Article	Sources	Article	Sources	Article
IEEE access	33	IEEE access	147	IEEE access	330
IEEE transactions on communications	8	IEEE transactions on wireless communications	44	IEEE open journal of the communications society	17
IEEE transactions on wireless communications	8	IEEE transactions on vehicular technology	20	IEEE transactions on wireless communications	7
IEEE journal on selected areas in communications	6	IEEE transactions on communications	18	IEEE internet of things journal	4
IEEE transactions on signal processing	5	IEEE internet of things journal	17	IEEE transactions on vehicular technology	4
IEEE transactions on vehicular technology	5	IEEE journal on selected areas in communications	14	Journal of communications and networks	4
IEEE transactions on mobile computing	3	IEEE open journal of the communications society	9	IEEE open journal of vehicular technology	3
Sensors (Switzerland)	3	IEEE wireless communications letters	9	IEEE systems journal	2
ETRI journal	2	IEEE communications surveys and tutorials	7	IEEE transactions on communications	2
IEEE wireless communications	2	IEEE communications magazine	4	Intelligent and converged networks	2

3.2.2. Most Locally Cited Authors, Articles Fractionalized: AF

The phrase “most locally cited authors” refers to authors cited the most within a specific collection of documents [46]. In the context of this review, it measures how frequently an author in power control techniques has received citations from other authors within the same domain [38]. Table 8 shows the most cited authors from each database. Authors whose work has significantly impacted power control techniques are Chen M and Karakiannidis GK, Poor HV, and Liu Y, respectively, from the Scopus, WoS, and IEEE databases. As observed, the IEEE database has the highest record of most local citations per author. It is observed that Liu Y has the highest score of citations.

Another parameter shown in Table 8 is the Articles Fractionalized (AF). By way of definition, fractionalization, in simple terms, is the method of dividing or breaking something into smaller parts or fractions [47]. It often refers to splitting a whole into smaller components or portions. For instance, in the context of AF, the research impact is divided among collaborating authors within scientific publications. It helps researchers to analyze fractionalization trends to understand collaboration patterns and the distribution of impact across different contributors [48,49] AF evaluates the productivity and the impact of authors, research, or academic institutions. It shows the percentage of articles that a certain author or organization contributed to the total number of articles in each dataset. For instance, author, Chen M had a significant contribution to the field with the highest value of 0.74 in the Scopus database.

Table 8. Most locally cited authors. Articles Fractionalized: AF.

Scopus			Web of Science			IEEE
Authors	Articles	AF	Authors	Articles	AF	Authors	Articles	AF
Chen M [50]	4	0.74	Poor HV [51]	6	1.14	Liu Y [52]	10	2.23
Karagiannidis GK [53]	4	0.63	Ding ZG [54]	5	1.01	Zhang Y [55]	7	1.92
Liu Y [52]	3	0.87	Han Z [56]	5	0.7	Li J [57]	6	1.09
Pan C [58]	3	0.62	Tafazolli R [59]	5	0.87	Tafazolli R [59]	6	1.12
Pan Y [60]	3	0.62	Zhang HJ [61]	5	0.96	Liu X [62]	6	1.22
Wang Y [63]	3	0.65	Zhang L [64]	5	1.78	Zhang X [65]	6	1.50
Yang Z [66]	3	0.62	Duong TQ [67]	4	0.87	Li Y [68]	6	1.13
Zhang H [69]	3	0.56	Liang YC [63]	4	1.2	Han Z [56]	6	1.21
Boudreau G [70]	2	0.40	Long KP [71]	4	0.83	Liu Z [72]	6	1.11
Chen G [73]	2	0.67	Shikh-Bahaei M [74]	4	0.85	Zhang Z [75]	6	1.57

Table 8. Most locally cited authors. Articles Fractionalized: AF.

Scopus			Web of Science			IEEE
Authors	Articles	AF	Authors	Articles	AF	Authors	Articles	AF
Chen M [50]	4	0.74	Poor HV [51]	6	1.14	Liu Y [52]	10	2.23
Karagiannidis GK [53]	4	0.63	Ding ZG [54]	5	1.01	Zhang Y [55]	7	1.92
Liu Y [52]	3	0.87	Han Z [56]	5	0.7	Li J [57]	6	1.09
Pan C [58]	3	0.62	Tafazolli R [59]	5	0.87	Tafazolli R [59]	6	1.12
Pan Y [60]	3	0.62	Zhang HJ [61]	5	0.96	Liu X [62]	6	1.22
Wang Y [63]	3	0.65	Zhang L [64]	5	1.78	Zhang X [65]	6	1.50
Yang Z [66]	3	0.62	Duong TQ [67]	4	0.87	Li Y [68]	6	1.13
Zhang H [69]	3	0.56	Liang YC [63]	4	1.2	Han Z [56]	6	1.21
Boudreau G [70]	2	0.40	Long KP [71]	4	0.83	Liu Z [72]	6	1.11
Chen G [73]	2	0.67	Shikh-Bahaei M [74]	4	0.85	Zhang Z [75]	6	1.57

3.3. Documents

3.3.1. Retrieved Documents

Table 9 represents the twenty-five most widely cited documents from the Scopus database on power control from 2018 to 2023. The article by Nasir et al. [76] is the most globally cited document, with 309 citations and higher Total Citations per year (61.8), published in IEEE journal on selected areas in communications.

Table 9. Scopus—most globally cited documents. DOI: Digital Object Identifier; TC: Total Citations; TCpY: TC per year; and NTC: Normalized TC.

Author	Year	Source	DOI	TC	TCpY	NTC
Nasir YS [76]	2019	IEEE journal on selected areas in communications	10.1109/JSAC.2019.2933973	309	61.80	5.76
Sun Y [77]	2019	IEEE communications surveys and tutorials	10.1109/COMST.2019.2924243	297	59.40	5.53
Naqvi SAR [78]	2018	IEEE communications magazine	10.1109/MCOM.2017.1700451	195	32.50	5.34
Pan Y [60]	2018	IEEE wireless communications letters	10.1109/LWC.2017.2759114	113	18.83	3.10
Zhang H [69]	2020	IEEE transactions on wireless communications	10.1109/TWC.2019.2948874	87	21.75	4.05
Shen Y [79]	2021	IEEE journal on selected areas in communications	10.1109/JSAC.2020.3036965	69	23.00	4.89
Khan WU [80]	2020	IEEE transactions on vehicular technology	10.1109/TVT.2020.3038387	66	16.50	3.07
Vo NS [81]	2018	IEEE access	10.1109/ACCESS.2017.2761978	65	10.83	1.78
Yang Z [82]	2018	IEEE transactions on wireless communications	10.1109/TWC.2017.2772824	61	10.17	1.67
Zhou T [83]	2018	IEEE transactions on vehicular technology	10.1109/TVT.2017.2768574	58	9.67	1.59
Niu Y [84]	2018	IEEE transactions on communications	10.1109/TCOMM.2017.2773529	53	8.83	1.45
Chincoli M [85]	2018	Sensors	10.3390/s18020375	51	8.50	1.40
Yoon T [86]	2018	IEEE access	10.1109/ACCESS.2018.2876354	44	7.33	1.21
Selim MM [87]	2019	IEEE access	10.1109/ACCESS.2019.2894390	34	6.80	0.63
Xiao J [88]	2018	IEEE transactions on communications	10.1109/TCOMM.2018.2851215	31	5.17	0.85
Girmay GG [89]	2019	IEEE access	10.1109/ACCESS.2019.2898645	30	6.00	0.56
Giannopoulos A [90]	2021	IEEE access	10.1109/ACCESS.2021.3113501	28	9.33	1.98
Yang Z [66]	2018	IEEE transactions on communications	10.1109/TCOMM.2018.2807443	25	4.17	0.68
Madani N [91]	2018	IEEE access	10.1109/ACCESS.2018.2855753	25	4.17	0.68
Sboui L [92]	2019	IEEE wireless communications	10.1109/MWC.2019.1800161	24	4.80	0.45
Elhattab M [93]	2022	IEEE transactions on wireless communications	10.1109/TWC.2021.3120989	23	11.50	4.24
Xu C, 2020 [94]	2020	IEEE access	10.1109/ACCESS.2020.2984798	23	5.75	1.07
Iqbal J, 2019 [95]	2019	IEEE access	10.1109/ACCESS.2019.2914486	20	4.00	0.37
Wang Y, 2020 [63]	2020	IEEE transactions on vehicular technology	10.1109/TVT.2020.3012311	20	5.00	0.93
Liu L, 2019 [96]	2019	IEEE transactions on mobile computing	10.1109/TMC.2018.2861859	16	3.20	0.30

Table 10 represents the twenty-five most globally cited documents retrieved from WoS on power control from 2018 to 2023. The article by Yang et al. [97].is the most globally cited document, with 268 citations and higher Total Citations per year (89.33), published in IEEE transactions on wireless communications.

Table 10. WoS—most globally cited documents. TC: Total Citations; TCpY: TC per year; and NTC: Normalized TC.

Paper	Year	Source	DOI	TC	TCpY	NTC
Yang ZH [97]	2021	IEEE transactions on wireless communications	10.1109/TWC.2020.3037554	268	89.33	18.52
Nasir YS [76]	2019	IEEE journal on selected areas in communications	10.1109/JSAC.2019.2933973	259	51.80	5.69
Sun YH [77]	2019	IEEE communications surveys and tutorials	10.1109/COMST.2019.2924243	252	50.40	5.53
Vaezi M [98]	2019	IEEE transactions on wireless communications	10.1109/MWC.2019.1800598	141	28.20	3.10
Du JB [99]	2020	IEEE internet of things journal	10.1109/JIOT.2020.3003449	130	32.50	5.44
Chowdhury MZ [100]	2020	IEEE communications surveys and tutorials	10.1109/COMST.2020.2966855	125	31.25	5.23
Zhang L [64]	2019	IEEE wireless communications	10.1109/MWC.2018.1800259	118	23.60	2.59
Moura J [101]	2019	IEEE communications surveys and tutorials	10.1109/COMST.2018.2863030	114	22.80	2.50
Rinaldi F [102]	2020	IEEE access	10.1109/ACCESS.2020.3022981	108	27.00	4.52
Zhang XR [103]	2020	IEEE internet of things journal	10.1109/JIOT.2019.2962715	104	26.00	4.35
Liu CX [104]	2021	IEEE journal on selected areas in communications	10.1109/JSAC.2020.3018837	88	29.33	6.08
Bonati L [105]	2020	Computer Networks	10.1016/j.comnet.2020.107516	87	21.75	3.64
Pocovi G [106]	2018	IEEE access	10.1109/ACCESS.2018.2838585	82	13.67	2.65
Meng F [107]	2020	IEEE transactions on wireless communications	10.1109/TWC.2020.3001736	82	20.50	3.43
Zhang HJ [69]	2020	IEEE transactions on wireless communications	10.1109/TWC.2019.2948874	81	20.25	3.39
Nwankwo CD [108]	2018	IEEE access	10.1109/ACCESS.2017.2774143	74	12.33	2.40
Liu DZ [109]	2021	IEEE journal on selected areas in communications	10.1109/JSAC.2020.3036948	72	24.00	4.98
Lagen S [110]	2020	IEEE communications surveys and tutorials	10.1109/COMST.2019.2949145	67	16.75	2.80
Pan YJ [111]	2021	IEEE wireless communications letters	10.1109/LWC.2021.3063365	66	22.00	4.56
Zhang WS [112]	2018	IEEE transactions on communications	10.1109/TCOMM.2018.2863385	64	10.67	2.07
Esswie AA [113]	2018	IEEE access	10.1109/ACCESS.2018.2854292	63	10.50	2.04
Khan WU [80]	2020	IEEE transactions on vehicular technology	10.1109/TVT.2020.3038387	61	15.25	2.55
Vo NS [81]	2018	IEEE access	10.1109/ACCESS.2017.2761978	59	9.83	1.91
Zhang HJ [61]	2020	IEEE journal on selected areas in communications	10.1109/JSAC.2020.3005489	59	14.75	2.47
Luong NC [114]	2019	IEEE communications surveys and tutorials	10.1109/COMST.2018.2870996	58	11.60	1.27

Table 11 represents the twenty-five most globally cited documents, retrieved from the IEEE Xplore database, on power control from 2018 to 2023. The article by Tayyab et al. [115] is the most globally cited document, with 309 citations and higher Total Citations per year (61.8), published in IEEE Access.

Table 11. IEEE—most globally cited documents. TC: Total Citation; TCpY: TC per year; and NTC: Normalized TC.

Paper	Year	Source	DOI	TC	TCpY	NTC
Tayyab M [115]	2019	IEEE access	10.1109/ACCESS.2019.2937405	112	22.40	5.72
Zhu L [116]	2019	IEEE access	10.1109/ACCESS.2019.2935169	104	20.80	5.31
Pocovi G [106]	2018	IEEE access	10.1109/ACCESS.2018.2838585	98	16.33	3.56
Bazzi A [117]	2018	IEEE access	10.1109/ACCESS.2018.2883401	98	16.33	3.56
Omri A [118]	2019	IEEE access	10.1109/ACCESS.2019.2907970	87	17.40	4.44
Adedoyin MA [119]	2020	IEEE access	10.1109/ACCESS.2020.2969980	80	20.00	4.88
Naik G [120]	2020	IEEE access	10.1109/ACCESS.2020.3016036	78	19.50	4.75
Nwankwo CD [108]	2018	IEEE access	10.1109/ACCESS.2017.2774143	73	12.17	2.65
Esswie AA [113]	2018	IEEE access	10.1109/ACCESS.2018.2854292	73	12.17	2.65
Budhiraja I [121]	2021	IEEE access	10.1109/ACCESS.2021.3081601	70	23.33	6.78
Vo NS [81]	2018	IEEE access	10.1109/ACCESS.2017.2761978	65	10.83	2.36
Li Z [122]	2019	IEEE access	10.1109/ACCESS.2019.2945478	65	13.00	3.32
Zhao J [123]	2018	IEEE access	10.1109/ACCESS.2018.2794770	61	10.17	2.21
Omar MS [124]	2018	IEEE internet of things journal	10.1109/JIOT.2017.2788362	61	10.17	2.21
Yang Z [82]	2018	IEEE transactions on wireless communications	10.1109/TWC.2017.2772824	59	9.83	2.14
Park J [125]	2019	IEEE access	10.1109/ACCESS.2019.2941278	59	11.80	3.01
Zhang S [126]	2018	IEEE access	10.1109/ACCESS.2018.2873804	58	9.67	2.11
Ma B [127]	2020	IEEE access	10.1109/ACCESS.2020.2975004	57	14.25	3.47
Moltafet A [128]	2018	IEEE access	10.1109/ACCESS.2017.2775338	57	9.50	2.07
Yazar A [129]	2018	IEEE access	10.1109/ACCESS.2018.2795752	54	9.00	1.96
Ma W [130]	2019	IEEE access	10.1109/ACCESS.2019.2921316	52	10.40	2.65
Alemaishat S [131]	2019	IEEE access	10.1109/ACCESS.2019.2937401	51	10.20	2.60
Bairagi AK [132]	2018	IEEE access	10.1109/ACCESS.2018.2868875	50	8.33	1.82
Alhammadi A [133]	2020	IEEE access	10.1109/ACCESS.2019.2961186	49	12.25	2.99
Li J [57]	2019	IEEE access	10.1109/ACCESS.2019.2900422	46	9.20	2.35

3.3.2. Summary of Retrieved Articles

The review retrieved the most globally cited articles from the Scopus, WoS, and IEEE databases using R package software. The most globally cited documents were screened to include relevant articles for further analysis. The full texts were investigated to identify relevant articles. A total of 32 articles were eventually included for further analysis. We classified the articles based on the power control taxonomy and the technique(s) employed in interference challenges in 5G networks. Articles not addressing power control technique(s) were not considered for further analysis. Furthermore, we provided additional details on each article’s contribution, limitations, or future work. Table 12 presents a summary of the included articles.

Table 12. A summary of articles on power control techniques.

Source	Year	Taxonomy (See Code) *	Technique (See Code) **	Contributions	Limitations/Future Work
Nasir and Guo [76]	2019	ML	DL	Achieves near-optimal power allocation in real time.	Future work may investigate better and easily tunable training and exploration schemes that can adapt to the environment non-stationarity of the multi-agent setting.
Sun et al. [77]	2019	ML	MLT	Provides guidance on which ML technique to implement, e.g., power control.	Further study is required to implement ML-based paradigms using open datasets and environments.
Pan et al. [60]	2018	DPC	DBIA	Maximizes the sum rate of D2D pairs while guaranteeing the minimum rate requirements of NOMA-based cellular users.	Co-channel interference occurs during SIC decoding when D2D pairs reuse spectrum with NOMA-based cellular users. To overcome this, additional power control and channel restrictions should be imposed (recommended for further study).
Zhang et al. [69]	2020	OPT/LP	LPOA	Addresses energy consumption in wireless networks.	Future work will extend to study nonlinear energy harvesting model(s) and consider more cases and techniques, such as massive MIMO situations.
Shen et al. [79]	2021	DL/OPT	GNNs	A scalable learning-based method is developed to solve radio resource management problems in dense wireless networks.	Future work may investigate the distributed deployment of message-passing graph neural networks (MPGNNs) for radio resource management in wireless networks and extend the theoretical results to more general application scenarios.
Khan et al. [80]	2020	OPT/LP	SQP	Addresses power allocation and spectrum efficiency.	The existing work is limited to two NOMA user scenarios. Allocating a single frequency block to only two user cases may not be enough to connect many IoT devices in the network. However, when many NOMA-enabled IoT devices are connected, interference occurs.
Vo et al. [81]	2018	H	GA	Achieves high-performance video streaming. Achieves high QoE at high energy efficiency while limiting co-channel interference over D2D communications in 5G networks.	The fluctuation constraint has not been considered in the optimization problem to reduce the complexity of GAs.The application of the GA could be extended to 5G and beyond to evaluate similar performance measures.
Yang et al. [82]	2018	OPT/LP	DPC	Finds an optimal power allocation strategy for multi-cell networks.	The users’ mobility issue for multi-cell NOMA systems is left for future work.
Zhou et al. [83]	2018	H	LBPC	Develops a load balancing method to perform user association and power updates for each base station.	Future work may consider load balancing in uplink heterogeneous cellular networks.
Niu et al. [84]	2018	H	EMS	Achieves high energy efficiency between D2D communications and concurrent.	Future work may evaluate the performance of the proposed scheme on a test bed using practical phased-array antennas and investigate energy consumption in pairwise D2D locations.
Chincoli and Antonio [85]	2018	ML	MLT	Lowers transmission power at the nodes to reduce energy consumption and interference.	This work could be extended to perform a more extensive comparative analysis to identify any deficiencies in the simulator.
Yoon et al. [86]	2018	OPT/LP	MILP	Optimizes the total transmit power for all users in a wireless network.	The proposed algorithms could be deployed to address power control challenges in small cell networks such as picocells and femtocells.
Selim et al. [87]	2019	OPT/LP	PCA	Addresses the problem of interference management between deployed UAVs and underlaying heterogeneous networks.	A comparative study could investigate outage probability and power control of D2D underlaying both OMA and NOMA UAV-assisted networks.
Xiao and Yang [88]	2018	OPT/LP	JIM	Reduces interference from other domains, balancing costs and benefits.	The study analyzed interference among cooperative small cells. Future studies may also consider interference management among non-cooperative small cells.
Girmay and Pham [89]	2019	OPT/LP	PSO	Guarantees the minimum data rate requirement of WiFi users by managing interference to improve the overall throughput of both cellular and D2D users.	The work could be extended to deploy other optimization techniques and perform a comparative study.
Yang et al. [66]	2018	OPT/LP	DPC	Finds an optimal power allocation strategy using a distributed algorithm for multi-cell networks.	Future work may consider synchronizing the DL and UL slots and the users’ non-ideal energy harvesting characteristics.
Madani [91]	2018	OPT/LP	DPC	Achieves a higher sum rate and individual rates in NOMA than OMA.	Small cells could be incorporated to analyze the performance of both NOMA and OMA.
Sboui et al. [92]	2019	OPT/LP	DPC	Identifies a power control scheme that can increase spectral and energy efficiency with no trade-off.	Since HetNets are considered one of the enabling technologies of 5G that promotes more energy-efficient communications, it may be appealing to study the implications of using several radio interfaces on energy efficiency in multi-radio IoT/user applications.
Elhattab et al. [93]	2022	OPT/LP	DPC	Minimizes the total transmit power at both the BS and the user-cooperating relay for each relaying mode.	Owing to the advances in cloud computing and the introduction of a cloud radio access network (C-RAN), the run time should be reduced to ease computational complexity.
Xu et al. [94]	2020	OPT/LP	DPC	Reduces total energy consumption in a novel NOMA-based mobile edge computing (MEC) network.	Future work may investigate the offloading decision scenario in NOMA-based MEC networks, where whether computing tasks can be offloaded to MEC servers or executed locally can be decided.
Wang et al. [63]	2020	OPT/LP	DPC	Minimizes interference, reduces power consumption, and improves network performance while guaranteeing UEs’ quality of service.	Due to the complexity of the scenarios under investigation, future studies may consider developing strategies to reduce the computational complexity in a nonlinear approximation of the problem.
Liu et al. [96]	2019	OPT/LP	DPC	Reduces total energy consumption in a novel NOMA-based MEC network.	Future work may consider adopting the concept of transmission patterns in online scheduling algorithms for joint scheduling and power allocation in a dynamic network and combine the decomposition approach with distributed algorithms to reduce the complexity.
Du et al. [99]	2020	OPT/LP	DL	Minimizes the long-term energy consumption of a THz wireless access-based MEC system to support high-quality immersive VR video services.	Further study may involve the application of DRL for further performance improvement in problem solving. The DRL technique could be applied to 5G and beyond networks to evaluate similar performance measures.
Zhang et al. [103]	2020	OPT/LP	DL	Maximizes the sum capacity of vehicle-to-infrastructure users while meeting the latency and reliability requirements of V2V pairs.	The application of the GA could be extended to 5G and beyond to evaluate similar performance measures.
Meng et al. [107]	2020	DL	DL	Applies DRL algorithms to address static optimization problems such as the centralized dynamic power allocation in multi-user cellular networks.	The actor-critic deep deterministic policy gradient (DDPG) algorithm can be applied to many problems, such as user scheduling, channel management, and power allocation in various communication networks.
Zhang et al. [69]	2020	OPT/LP	DPC	Maximizes energy efficiency with a power optimization algorithm using Lagrangian duality.	Future work may study a more realistic (nonlinear) energy harvesting model and consider more cases and techniques, such as massive MIMO for multi-antenna situations.
Liu and Simeone [109]	2021	H	DPC	Minimizes the learning optimality gap under privacy and power constraints.	A possible extension would be to study the optimization of the threshold for channel inversion to maximize the learning performance under privacy and power constraints. The study could be generalized to other network topologies, including multi-hop (D2D) networks. Another direction is to consider the implementation of digital transmission, where quantization introduces an additional privacy-preserving mechanism on top of the channel noise.
Li et al. [122]	2019	H	IOP	IOP schemes can realize the alignment of multiple interferences from the same source with only one degree of freedom cost at the interfered receiver.	Future studies will seek to conduct a detailed analysis of the complexity of Inside-Out Precoding (IOP) and design them with less complexity and/or cost.
Zhang and Zhao [126]	2018	OPT/LP	DPC	Both primary and secondary users share the same licenced spectrum.	A further in-depth topology-aware study is proposed in the more generic setting of multi-radio, multi-channel wireless networks.
Yazar [129]	2018	H	MN	Determines the efficient number of numerologies to inter-numerology interference, spectral efficiency, complexity, and signalling overhead.	Future studies may include (1) the development of optimization algorithms for non-orthogonal networks, (2) the alignment of mixed numerology structures with adaptive guard bands between consecutive numerologies to enhance the spectral efficiency of the system, and (3) employing different waveforms with different numerology parameters for 5G and beyond networks.
Alemaishat et al. [131]	2019	DPC	KMT	Improved energy efficiency and throughput are realized.	Future work may involve the evaluation of the influence of different numbers of D2D receivers.

Taxonomy (Linear programming: LP; Dynamic power control: DPC; DL; machine learning: ML; optimization: OPT; heuristic: H) *. Technique (deep learning: DL; particle swarm optimization: PSO; machine learning technique: MLT; graph neural networks: GNNs; distributed joint interference management: JIM; optimal sequential quadratic programming: SQP; Langrangian power optimization algorithm: LPOA; dual-based iterative algorithm: DBIA; Kuhn–Munkres technique: KMT; Inside-Out Precoding: IOP; Genetic Algorithm: GA; load balancing power control: LBPC; Energy-Efficient Multicast Scheduling: EMS; mixed-integer linear program: MILP; power control algorithm: PCA; mixed numerology: MN) **.

3.3.3. Power Control Technique Occurrences

Figure 4 shows the number of occurrences of power control techniques from the summary of the included articles in Table 12. The figure shows the methods and their corresponding occurrences (also referred to as count in the figure). The results show that dynamic power control (DPC) has the highest count. In bibliometric analysis, a high occurrence of an element under study, such as the author’s keywords and number of citations, is typically recognized as significant. Thus, this analysis concludes that the research hotspot is in dynamic power control. Likewise, we recommend further research in this area of study.

Figure 4. Power control techniques obtained from the databases.

4. Power Control Technique Evaluation

4.1. Power Control Theories

The main power control theories are classified according to their specific applications. The classification encompasses mathematical programming theory, neural networks, game theory, deep learning, dynamic control theory, machine learning, and network protocol heuristics, as shown in Figure 5.

Figure 5. Classification of power control techniques. Adapted from [17].

Most power control problems in wireless networks can be formulated as optimization problems by applying linear programming theory [7]. For instance, maximizing system throughput is subject to QoS requirements from users [8]. In a complex application where a system is dynamic, like in a multi-hop wireless system, an optimization problem can be an NP-hard problem [9]. Network protocol heuristics are applied to reduce the complexity of a dynamic power control theory (DPCT) problem [10]. Some of the power optimization problems are focused on solving the SINR. For instance, in [11], a DPCT algorithm is proposed to maximize the SINR subject to maximum power constraints. The algorithms proposed in [12,13] minimize power consumption in the event of large-scale fading. Another power control application is found in [14], where total power transmitted is minimized over discrete power levels. In [15,16], the SIR outage probability is minimized by total transmission power.

In [18], the authors developed an efficient graph neural network (GNN) architecture that is scalable and generates the same output for the transformed input function. Their algorithm addresses interference challenges amongst multiple access points (APs) and a UE group. Compared to the baseline algorithms, the technique verifies that adaptability to minimum-capacity constraints of network configurations and channel conditions is achieved.

The authors in [79] proposed applying the graph neural network (GNN) technique to address radio resource management challenges in wireless networks. A comparative study of this technique was performed with the family of distributed optimization algorithms. The method was performed to simulate power control and beamforming. The results showed that a GNN performs better than traditional optimization-based techniques without domain knowledge. Furthermore, it has been demonstrated that a GNN satisfies the permutation equivariance property [134,135], can generalize large-scale problems [136], and improve high computational efficiency [137].

The application of game theory in wireless networks is commonplace. Game theoretic power control techniques can be modelled using cooperative or non-cooperative approaches. These approaches are extensively discussed in [138,139]. For instance, the authors in [140] combined successive interference cancellation (SIC) algorithms [141] and non-cooperative game theory [138] approaches to save energy and improve network throughput. They presented a non-cooperative game theory power control approach based on CDMA wireless sensor networks (WSNs) in their groundbreaking study. Sensor node energy consumption is reduced, and network lifetime is increased when this method is used in conjunction with other power control techniques in WSNs.

Mobile operators deploy more small cells to support users with varied data rate requirements in 5G mobile communication systems. Microcell users attached to small cells use the same frequency band to enhance capacity. As a result of this deployment, co-channel interference is likely to occur, which in turn causes poor network performance. To address this challenge, the authors of [142] proposed a game theory-based power control technique to optimize femtocell user power. The technique improves the sum rate and reduces outage and interference compared to the typical power control schemes.

A power control strategy based on game theory was proposed in [143] to remove interference between D2D lines and cellular user devices. The benefits are reliable connectivity and minimal power consumption in wireless communication networks. In [144], an evolutionary game theory model is proposed to adaptively adjust the transmitted power level of the users to mitigate interference among users in 5G networks. The proposed technique achieves higher efficiency regarding network spectral and energy efficiencies. Furthermore, the authors in [145] modelled non-cooperative power control in WCDMA and OFDMA systems as evolutionary games to study interference among a dense network of mobiles.

Machine learning [146] techniques are suitable in a dynamic wireless environment. When applied with deep learning techniques, wireless networks can perform intelligent operations such as finding hotspots, locating interference distribution, and identifying traffic congestion points [146]. These techniques are widely discussed in [147,148]. For instance, ref. [149] used a multi-agent deep reinforcement learning approach to implement distributed resource management and interference mitigation mechanisms in wireless networks. The simulation results show that the suggested strategy is superior to decentralized models. The results were compared in terms of the average and fifth-percentile user rates.

4.2. Power Control Techniques: Case Studies

Several case studies on power control techniques highlight their practical applications and benefits in 5G networks. This study presents a collection of case studies with real power control applications in interference management. The case studies outline strategies to address challenges, practical suggestions, or novel solutions. The following case studies illustrate the application of power control techniques in 5G networks to address interference.

4.2.1. Case Study 1: METIS-II Perspective

Ultra-dense networks in 5G NR are deployed to address the growing wireless data traffic challenge. However, when data traffic increases, interference problems arise, and the user’s network performance is affected.

The study discusses various interference management drivers to improve network performance. The survey focuses on enhancing cell-edge throughput, reducing energy consumption, and minimizing signalling overhead.

The study explores techniques such as beamforming, dynamic time-division duplex (TDD), and self-backhauling.

Details of the case study are found in [150].

4.2.2. Case Study 2: Interference Management in 5G and Beyond Networks

A sudden increase in wireless data services has been observed over the past ten years. Data traffic has been growing at a high rate. Therefore, fifth generation (5G) and future mobile communication services should improve technologies to manage this level of growth, for instance, through the use of terahertz communications and the introduction of non-terrestrial networks. On the one hand, there is technological improvement, but on the other hand, there is an interference challenge that needs to be managed properly. To address this challenge, the authors provided a survey of interference management in 5G and beyond networks and discuss its future evolution.

This survey examines cross-link, multi-user, and inter-cell interference management strategies. It also covers network densification, massive MIMO, and the use of millimetre-wave spectrum.

The survey classifies and explains different types of interference. This is followed by presenting a taxonomy of existing interference management approaches. The survey explains interference measurement reports and signalling in 5G NR. The survey presents an in-depth literature review and technical discussion of appropriate management schemes for each type of interference identified. The survey concludes by discussing the main interference challenges encountered in future 6G networks and presenting insights on the suggested new interference management approaches, including useful guidelines for an AI-based solution.

This review provides a first-hand guide to the industry in determining the most relevant technology for interference management. It also allows for the consideration of future challenges and research directions.

Details of the case study are found in [151].

4.2.3. Case Study 3: Interference Management Strategy for 5G Femtocell Clusters

This review paper examines the concerns of interferences discovered and researched in various network topologies and techniques. It also pays attention to recent advances in its management, such as advanced receiver, joint scheduling, and network information theory, among others. Potential advantages of each interference management technique are illustrated, and it is proven that if 5G cellular networks use intricate joint scheduling, the advantages of sophisticated receivers can be effectively utilized.

This study explores interference management strategies for femtocell clusters, addressing inter-cell interference and power control challenges.

Details of the case study are found in [13].

4.3. Keywords Analysis

Keywords play a role in shedding light on the status of the research topics in a field. This section presents the bibliometric keyword analysis of the articles published in the Scopus, WoS, and IEEE databases from 2018 to 2023.

(i)

Most frequent keywords

The author’s keywords generated a total of 2788 occurrences. The occurrences from the Scopus, WoS, and IEEE databases were 181, 1305, and 1302, respectively. For analysis, we selected the first 50 articles from each database. Synonymous keywords were combined. Table 13 shows some of the keywords with similar meanings. Keywords that seemed to have no bearing in the study were discarded. Those that were removed included keywords such as “fairness,” “scheduling”, “relays”, and servers.

Table 13. Keywords with the same meaning.

Keyword	Combined Keyword(s)
resource management	radio resource management
interference management	interference mitigation
5g	5g mobile communication
cellular network	cellular networks
quality of service	qos
non-orthogonal multiple access	noma
heterogeneous network	hetnets
millimetre wave	millimetre-wave

Table 14 lists the remaining keywords after combining synonyms and removing those that seemed irrelevant. The table shows that only 45 keywords were left for analysis. The table shows three parameters. The first column is the list of keywords. The second column shows the number of times the keywords were listed from some of the 770 articles. The third and fourth columns show the residuals and standardized residuals (SRs). To generate the values of the third and fourth columns, a one-sample t-test was performed using a statistic kingdom calculator. The calculator was set to a 0.05 significance level (α) and an expected mean of 44 (from the data). The residual was determined from the Pearson residual (1):

$$SR={\displaystyle \frac{y_i-{\widehat{\mu }}_i}{\sqrt{{\widehat{\mu }}_i}}}$$

(1)

where $y_i$ is the observed values, and ${\widehat{\mu }}_i$ is the expected mean.

Table 14. Noteworthy keywords depicting frequency of occurrences, percentages, and residuals.

Keywords	Occurrences (%)	Residual	Standardized Residual
resource management/radio resource management	274 (35.58%)	34.67	28.56
power control	173 (22.47%)	19.45	16.02
non-orthogonal multiple access/noma	157 (20.39%)	17.04	14.03
interference	127 (16.49%)	12.51	10.31
5g/5g mobile communication	115 (14.94%)	10.70	8.82
optimization/nonconvex optimization	113 (14.68%)	10.40	8.57
resource allocation	106 (13.77%)	9.35	7.70
device-to-device communication/device-to-device/d2d	76 (9.87%)	4.82	3.97
power allocation	62 (8.05%)	2.71	2.24
quality of service/qos	59 (7.66%)	2.26	1.86
interference management/interference mitigation	50 (6.49%)	0.90	0.75
reinforcement learning/deep reinforcement learning	42 (5.45%)	−0.30	−0.25
reinforcement learning	40 (5.19%)	−0.60	−0.50
6g mobile communication/6g	38 (4.94%)	−0.90	−0.75
downlink	37 (4.81%)	−1.06	−0.87
uplink	34 (4.42%)	−1.51	−1.24
internet of things (iot)	32 (4.16%)	−1.81	−1.49
heterogeneous cellular networks	30 (3.90%)	−2.11	−1.74
throughput	30 (3.90%)	−2.11	−1.74
machine learning	29 (3.77%)	−2.26	−1.86
massive mimo	29 (3.77%)	−2.26	−1.86
cellular network	28 (3.64%)	−2.41	−1.99
signal to noise ratio	27 (3.51%)	−2.56	−2.11
spectral efficiency	25 (3.25%)	−2.86	−2.36
game theory	24 (3.12%)	−3.02	−2.48
wireless networks	23 (2.99%)	−3.17	−2.61
handover	18(2.34%)	−3.92	−3.23
training	18 (2.34%)	−3.92	−3.23
energy efficiency	15 (1.95%)	−4.37	−3.60
heuristic algorithms	15 (1.95%)	−4.37	−3.60
fading channels	15 (1.95%)	−4.37	−3.60
ofdm	13 (1.69%)	−4.67	−3.85
small cells/femtocells	12(1.56%)	−4.82	−3.97
millimeter wave	12 (1.56%)	−4.82	−3.97
ultra reliable low latency communication	11 (1.43%)	−4.97	−4.10
stochastic geometry	11 (1.43%)	−4.97	−4.10
cognitive radio	11 (1.43%)	−4.97	−4.10
channel estimation	11 (1.43%)	−4.97	−4.10
deep learning/deep q-learning	9 (1.17%)	−5.28	−4.35
spectrum sharing	8 (1.04%)	−5.43	−4.47
mobile edge computing	5 (0.65%)	−5.88	−4.84
unsupervised learning	4 (0.52%)	−6.03	−4.97
graph neural networks	3 (0.39%)	−6.18	−5.09
linear programming	2 (0.26%)	−6.33	−5.22

The standardized residuals (SRs) are determined by (2) as follows:

$$SR={\displaystyle \frac{y_i-{\widehat{\mu }}_i}{SE}}$$

(2)

The residual values also emphasize the occurrence of keywords. The positive values indicate a high expected occurrence, and the negative values indicate a lower expected frequency. “Resource management” was the most frequently used keyword. It was listed as a keyword in 36% of the articles. Second on the list is the keyword “power control”, with 22% of occurrence.

Keywords identify the relevant documents by typing in the database’s search box. They should be properly selected to obtain the required documents. Table 15 lists the 12 keywords with the highest scores from each database. It is noted that “resource management” is the most widely used term in the list, notably from the WoS database. The keywords “power control” and “interference” were used 173 and 123 times, respectively. The keywords “optimization” and “5g mobile communication” were also high, with 111 and 110 cumulative occurrences, respectively. These results indicate the ongoing interest in resource management, specifically focusing on power control and optimization in 5G mobile communication. The synonymous keywords for “resource management” and “power control” are “resource allocation” and “power allocation”, respectively. These synonymous keywords each scored 106 and 62 occurrences. Figure 6 shows the cumulative occurrence of the keywords shown in Table 15.

Table 15. Authors’ keyword occurrences. CO: cumulative occurrences.

Keywords	Occurrences			Cumulative Occurrences
	Scopus	WoS	IEEE Xplore
power control	29	96	48	173
interference management	17	0	17	34
resource allocation	13	40	53	106
energy efficiency	12	31	48	91
(radio) resource management	13	231	9	253
5g mobile communication	7	45	58	110
device-to-device communication	11	32	11	54
non-orthogonal multiple access (noma)	11	56	24	91
power allocation	6	26	30	62
reinforcement learning	6	23	11	40
interference	5	109	13	123
optimization	5	92	14	111

Figure 6. Most frequently used cumulative keywords from the databases.

Table 16 shows groups of keywords with records from the ¹ Scopus, ² WoS, and ³ IEEE databases. From each database, we obtained the number of occurrences for further analysis.

Table 16. Authors’ keyword occurrences from the sources.

Keywords	Occurrences	Keywords	Occurrences	Keywords	Occurrences
1 NM	8	3 RM	53	2 SE	0
2 NM	56	1 RA	17	3 SE	22
3 NM	24	2 RA	23	1 EE	12
1 PC	80	3 RA	53	2 EE	31
2 PC	96	1 RL	15	3 EE	48
3 PC	48	2 RL	23	1 GT	3
1 QoS	19	3 RL	6	2 GT	14
2 QoS	46	1 DL	13	3 GT	7
3 QoS	0	2 DL	18	1 IOT	0
1 5G	24	3 DL	5	2 IOT	22
2 5G	0	1 OPT	9	3 IOT	5
3 5G	58	2 OPT	96	1 SC	2
1 RM	39	3 OPT	14	2 SC	0
2 RM	231	1 SE	3	3 SC	10
Keyword acronyms
non-orthogonal multiple access: NM		resource allocation: RA		energy efficiency: EE
power control: PC		reinforcement learning: RL		game theory: GT
quality of service: QoS		deep learning: DL		Internet of Things: IoT
5G mobile communication: 5G		optimization: OPT		small cells: SC
resource management: RM		spectral efficiency: SE

¹ (Keyword): Scopus; ² (Keyword): WoS; ³ (Keyword): IEEE.

The keywords displayed in Figure 7 are classified in terms of the underlying technology standard, concepts underlying management of resources, technology standard in use, models for implementing resource management, and network performance indicators. The keywords related to the technology standard are “5G” and “SC”. The keywords “RM” and “RA” are aligned with the management and allocation of network resources. Keywords specific to modelling are “PC”, “GT”, “RL”, “DL”, and “OPT”. The performance-related keywords are “EE”, “SE”, and “QoS”.

Figure 7. Keyword co-occurrence box and whiskers plots.

The box plot in Figure 7 gives us a basic idea of the distribution of co-occurrence data. The data are spread out if the box plot is relatively tall. The box plot of the keyword “RM” is relatively tall. In this study’s context, “RM” has a high occurrence concerning other keywords. By observation, the means and standard deviations of both “RM” and “PC” are high compared to other keywords, as confirmed by the summary statistics in Table 17.

Table 17. Summary statistics by groups—count, mean, standard deviation.

Keywords	Count	Mean	Standard Deviation
NM	3	29.3	24.4
PC	3	74.7	24.4
QoS	3	21.7	23.1
5G	3	27.3	29.1
RM	3	108.0	107.0
RA	3	31.0	19.3
RL	3	14.7	8.5
DL	3	12.0	6.6
OPT	3	39.7	48.9
SE	3	8.3	11.9
EE	3	30.3	18.0
GT	3	8.0	5.6
IOT	3	9.0	11.5
SC	3	4.0	5.3

(ii)

Trending keywords and topics

The review shows that researchers are interested in radio resource management and power control in wireless networks. This is confirmed by the pie chart in Figure 8. The frequencies of keywords “SC” and “GT” are relatively low, showing less research interest in this study area. For bibliometric analysis, high-frequency keywords are typically recognized as significant research themes.

Figure 8. Keyword co-occurrence pie chart.

In addition to the keywords previously discussed above, the occurrence of keywords can be analyzed to reflect the central ideas of a research area and the trending topic in the research field. As discussed previously, “resource management” is seen as the most occurring keyword in this review. It is interesting to note that “power control” and “optimization” are still trending. There has been significant growth for the “resource management”, “power control”, and “optimization” keywords after 2018, indicating their continued relevance in 5G mobile communication. Several factors and trends inform the continued relevance of these keywords in wireless technology. These include the power control and optimization techniques’ role in reducing intra-cell and inter-cell interference and optimizing user equipment (UE) power consumption in mobile networks.

5. Discussion and Recommendations for Future Study

This paper has devoted much attention to reviewing the power control literature with a bias on interference management for wireless networks. The authors established that power control in wireless networks is carried out mainly to maintain an acceptable QoS [152,153], for example, in terms of the SINR or throughput for all devices by minimizing interferences among devices [154], and to optimize energy efficiency, particularly from the user’s perspective [155].

Extensive research work related to power control has been carried out. This paper retrieved 770 articles published from 2018 until 2023 from Scopus, WoS, and IEEE Xplore for review. To address the research questions, we obtained the (1) main information, (2) annual scientific production, (3) most locally cited authors, (4) most globally cited documents, (5) keyword co-occurrence, and trending topic parameters retrieved from the R package software. Even though the authors retrieved many articles, we understand that this review does not exhaust all the work covered in this area but rather highlights some of the contributions made to address interference management through power control techniques.

The number of publications published year-on-year provides a reasonable estimate of the current publication trend in a specific field of research. Analyzing the trend in the number of publications can provide information on the probable research trend in future. Table 5 shows, over the years, the number of publications and the distribution of the publications received on power control. The retrieved data show many articles from IEEE Xplore, followed by WoS. Scopus has the lowest record. This influenced the average co-authors’ score per document (3.89, lower than those of WoS and IEEE Xplore).

The journal with the most published papers was “IEEE access” in all databases. The highest records publications from Scopus, WoS, and IEEE Xplore were 33, 147, and 330, respectively. The second most popular sources were “IEEE transactions on communications”, “IEEE transactions on wireless communications”, and “IEEE open journal of the communications society” with records of 8, 44, and 17, respectively.

The “most locally cited authors” refers to authors cited the most within a specific collection of documents. In the context of this review, we measured how frequently an author in power control techniques has been cited by other authors. Authors whose work has significantly impacted power control techniques are Chen M and Karakiannidis GK, Poor HV, and Liu Y, respectively, from the Scopus, WoS, and IEEE databases. The IEEE database has the highest record of most local citations per author. We observed that Liu Y has the highest citation score.

Concerning keyword co-occurrence analysis, the results show that there has been significant growth for the “resource management,” “power control”, and “optimization” keywords after 2018, indicating their continued relevance in 5G mobile communication. Several factors and trends inform the continued relevance of these keywords in wireless technology. These include the power control and optimization techniques’ role in reducing intra-cell and inter-cell interference and optimizing user equipment (UE) power consumption in mobile networks.

The authors used the “most globally cited documents” from each database to obtain the occurrences of power control techniques. The results show that dynamic power control (DPC) has the highest count. In bibliometric analysis, the high occurrence of an element under study, such as the author’s keywords and number of citations, is typically recognized as significant. Thus, this analysis concludes that the research hotspot is in dynamic power control. Likewise, we recommend further research in this area of study.

Some recommended techniques for handling interference management and power control in wireless networks have been discussed. Even though some of these techniques are generally used in 4G series wireless networks, they can also be enhanced to be deployed in 5G wireless networks. As a common trend in wireless technology, there is always ongoing research to find the most optimal techniques to address interference in wireless networks. These techniques include mathematical programming theory, neural networks, game theory, deep learning, dynamic control theory, machine learning, and network protocol heuristics.

There has been significant growth for the “resource management,” “power control”, and “optimization” keywords after 2018, indicating their continued relevance in 5G mobile communication. Several factors and trends inform the continued relevance of these keywords in wireless technology. These include the power control and optimization techniques’ role in reducing intra-cell and inter-cell interference and optimizing user equipment (UE) power consumption in mobile networks.

6. Limitations

The study performed cannot be without limitations. The range of years from 2018 to 2023 is a limitation. The range could have been extended to 2018, when 5G technology was in the process of conceptualization. However, we understand that the rigorous procedure of our systematic literature review reduced the likelihood that the omitted research would have contained information that would significantly affect our results. Nevertheless, our future review will take this into account. Although three significant databases were used for searching relevant studies, many un-indexed journals and publications from such journals might have been missed. We recommend considering other pertinent databases such as ScienceDirect and Google Scholar. No search string can be 100% perfect; therefore, false positive and negative results are always possible. Even though other expert contributions are related to selecting search keywords from information specialists (i.e., librarians), more academics in the field will be involved for quality results. Only English language papers were considered for this study, which may have biassed results toward English-speaking countries to some extent. Collaboration with other non-English speaking authors will be sought (in future) to obtain a more comprehensive analysis.

7. Conclusions

In this paper, the authors conducted a systematic literature review to answer four research questions. The methodology involved a bibliometric analysis of the literature published on power control in wireless networks. The main goal was to inform us of the current and future research status of power control taxonomy and its applications to addressing interference in 5G networks. A bibliometric analysis software from an R package was used to analyze (1) the current publication trends, (2) the most influential authors, countries, and journal articles, (3) the current state of collaboration, and (4) the areas involving power control that need additional study. The authors obtained both numerical and graphical visualizations. The visualization results make several contributions to the current literature. First, it informs researchers in this area to quantify and visualize the thematic evolution of power control in 5G networks. Secondly, it provides insight into the current state of the topic under study and predicts the direction of future research. Thirdly, it tracks the words of authors, the impact of papers, and the trajectory of research ideas by measuring citation counts in critical research databases and online sources. Finally, it investigates the recurring themes in a research field and identifies possible future research directions. Overall, the visualization results confirmed continuous growth in the study, meaning that there is a need for further exploration of power control techniques in wireless networks. Based on the results, there is a definite need to establish the impact of power control in 5G and beyond wireless networks.