Decision Model for DLT Applicability in Recycling Life Cycle Tracking

Abstract

Circular economy implementation is feasible through digital tracking of raw materials. The aim of this article is to analyze the possibilities for deploying blockchain or Ethereum as Distributed Ledger Technology (DLT) platforms. Various decision-selection algorithms were examined for choosing a suitable platform and network type (public, private). The article analyses the drawbacks and barriers to achieving a circular economy in accordance with sustainable development criteria. Based on the regulatory framework and operational projects, it is shown that digital technologies improve the efficiency of raw material processing and life cycle tracking. To build an effective network, a separator is proposed as a separate participant in addition to producers, processors, and consumers. Their functions are waste classification and sorting, and receiving and transmitting heterogeneous data streams while preserving the privacy of sensitive ones. Drawing on the specifics of circular economy implementation, selection criteria are proposed, including deployment cost, data immutability, energy consumption, need for new technology, and others. A superior decision support tree tailored to material recycling is proposed. Its advantage is fast decision-making and selection of an appropriate platform within the DLT landscape. The proposed conceptual framework is not a universal solution as it can be adapted to different recycling schemes, especially when the regulatory framework, technologies, materials, flows, scheme participants, stakeholders, or other factors change.

1. Introduction

In today’s consumer society, the concept of a circular economy is evolving in line with sustainable development criteria. The transition from a linear to a circular economy with the potential to improve the efficiency of resource use [1,2] is particularly important for society and business. Through the adoption of digital technologies, a zero-waste city is envisaged via proper and complete recycling of waste [1,2,3]. Recycling itself is one stage in a long chain typical of the circular economy, which closes back at the start of the same process. The main activities include collection, storage, transport, separation, processing, repackaging, reuse, incineration, and ultimate disposal of waste. To carry out these activities effectively and achieve a circular economy in line with sustainable development criteria, the regulatory framework [4,5,6,7] recommends implementing integrated, modern digital technologies. This is shown in Figure 1. Digitalization and improved efficiency are closely linked to the use of machine learning and artificial intelligence [8]. Another approach is to build systems based on cradle-to-cradle design, life cycle thinking, and closed-loop flows [1].

Material recycling not only helps improve the environmental situation, but also enables innovations in materials science and significantly improves the social behavior of responsible citizens [1]. Although implementing the technology is expensive, it can ultimately have a positive economic effect.

Unfortunately, waste continues to accumulate; China is one of the largest producers [9], while in Indonesia, 60% is organic waste, 14% plastics, 9% paper, 4.3% metal, 1.7% glass, 3.5% textile, and 5.5% rubber [10]. Various recycling methods exist, one of the most innovative being the production of biodegradable plastics [11]; however, this is not the focus of this article. The innovation proposed here is a model for assessing the rationale for introducing digital technologies to achieve transparency and encourage citizens to take part in waste collection. Only through clarity, reduced traceability costs, and a real economic benefit for participants can the idea of introducing blockchain (BC) in this sector be realized.

For a genuine circular economy, flows must be tracked by type and quantity. There are substantial differences in identifying and registering solid, liquid, and gaseous pollutants [12]. The idea is that AI algorithms, along with fuzzy matching and deep learning, can detect the source of pollution, but the goal is to achieve transparency—and that is possible with blockchain technology [12]. Efficiency improves when trained specialists are available, when there is policy support for all participant groups, and when new technologies are present at every stage of the circular economy process [1].

The aim of the current work is the traceability of municipal solid waste based on digital technologies, which is easier at the output stage when digital scales are available. At the input stage, this is more difficult since waste arrives as a mix of different components from many sources. The mix includes organic matter, plastics, metal, glass, leather, and other materials. Achieving the goals set requires diverse activities, such as proper allocation of waste bins, their servicing, routing of collection trucks, sorting of waste, and subsequent processing [3]. The objectives of introducing digital technologies to manage these activities also include environmental protection, reduced consumption of natural resources, higher GDP, and others. According to analysts [13], these activities are linked to economic, environmental, and social assessment parameters, but above all to substantial financial resources and analyses [14]. In line with sustainable development criteria, the Circular Economy Strategy 2020 [4] prioritizes digitization of flow tracking and has an undisputed positive environmental and social effect. Such benefits include acquisition of data for analysis, improved process controllability, higher levels of use of recycled materials, etc.

The economic rationale for deploying digital technologies is compelling. Evidence includes the model of introducing tokens to pay for collected waste. This leads to higher collection rates, subsequent improvement of environmental conditions, and the generation of revenue [15]. Literature also cites advantages like social inclusion, accountability of participants, traceability and reporting by producers, environmentally friendly relationships, and a transparent supply chain [15], as shown in Figure 2.

A simple example can demonstrate the rationale for investing in digitalization, which involves not only developing DLT but also implementing sensors and payment applications [16]. Building the system will require staff training and improving citizens’ competence. Introducing tokens (or other means of payment) in accordance with the regulatory framework will raise participants’ welfare. Last but not least, waste recycling will improve environmental conditions. All these factors underpin sustainable development—social, economic, and environmental improvements for citizens [17].

Despite the multifaceted nature of the collection task, IBM is working on a potentially blockchain-based solution called Plastic Bank [15,18], while another DLT platform, IOTA [19], has proposed digitized, traceable waste collection.

High cost and implementation difficulties could be reduced through proper logistics and routing, as well as by selecting criteria for evaluating a suitable digital platform. This requires defining barriers to blockchain deployment across the entire circular economy chain: consumer–processor–producer–consumer.

The article concludes with a proposed conceptual design aid tree based on research that favors a directed acyclic graph (DAG), which ensures that a decision is reached [20], synthesized after analysis of known non-standardized models. It has been found that in recent years, AI-based decision methods are entering all fields of science, including the need to apply DLT in line with its use case. This is also evidenced by the applicability of DLT for management without reliance on artificial intelligence [21]. The proposed conceptual decision support tree is not standardized and is not a graph; however, it reflects the specifics of recycling technology. It is not concerned with configuring offline and online channels of consumer behavior, which is outside the scope of this study [22].

This study is a conceptual model without empirical research. The proposed decision support tree is based on rules synthesized on the criteria of a literature base for choosing between databases or DLT according to the specifics of distributed networks. This model is primarily for decision support, not an empirical tool.

The proposed framework achieves accountability, visibility, and traceability of material flows. This directly impacts improving resource efficiency (SDG 12), sustainable cities and communities (SDG 11), responsible production (SDG 12.5), and climate benefits through reduced unnecessary transport and higher recycling yields (SDG 13) of the sustainable development criteria. Unfortunately, indicators for their determination have been proposed due to a lack of real data.

In the recycling industry, different material flows are mixed, and specific data is shared throughout the life cycle. This is a prerequisite for audit complexity and regulatory constraints. To address this issue, a thorough analysis of the overall decision support framework for using DLT is required to achieve the following:

• Application and selection of DLT to the recycling context;
• Definition of the functions of the separation center as a new participant;
• Clear rules for selecting DLT or database;
• Synthesis of an appropriate decision support tree for selecting the architecture.
• Introducing an application or tokens to reward participants or third parties.

The study fills the design aid framework for DLT selection and an adequate architecture for tracking compatible material streams in multi-actor coordination.

2. Materials and Methods

At this point, based on the literature, the main methods for selecting sharing technology are introduced, in connection with problems in the field of waste recycling.

The study is a conceptual framework for decision-making based on published rules in the literature. The rules are analyzed, and a decision support tree is synthesized for technical, managerial, and operational activities. This is not a proven empirical model, and everyone is free to adjust it according to the technological line of processing in the life cycle.

According to the literature, barriers to introducing blockchain in recycling fall into cognitive, technical, internal, and external categories [15]. The first group covers citizens’ unwillingness or inability to take part in the traceability chain and a lack of knowledge about how to use new technologies. Inaccurate sensors and technical systems that do not interact properly with the new platform are classified as technical barriers. Internal barriers remain, such as the firms’ own costs of the deployment of innovative products. Last among these are factors external to the company, including legal obstacles to introducing tokens, the network to be built, protection against token theft, or cultural barriers.

The current article classifies the main barriers to DLT adoption in any industrial or societal domain as regulatory, infrastructural, technological, social, financial, and subjective. In this domain, there are no major regulatory obstacles since the existing legislation supports choosing a suitable digital technology for flow tracking.

For Bulgaria, however, sufficiently modern infrastructure is not yet in place [23]. In this respect, barriers for maintenance and operation exist even at the global scale. One example is monitoring waste levels in bins with ultrasonic sensors and subsequent collection by smart trucks. Vandalism, improper allocation, and other issues are reported [3]. Another problem is identifying waste via Quick Response (QR) codes or Radio Frequency Identification (RFID) tags, which cannot be read when waste is broken into fragments—an issue linked to supply chain management according to GDPR [24]. Overcoming these problems requires digitization of the entire material life cycle: collection, separation, recycling, and production.

Technological issues include the continuous growth in the quantity [25,26] and diversity of materials produced, the lack of a digital identity, the use of more than one type of raw material in a single product [25], reduced quality of recycled material [23], inability to process certain newly developed plastics at the plant, or incompatible technologies [25]. The latter leads to the inability to recycle a large share of existing plastics [23] and motivates establishing—or treating as a separate participant—digital separation facilities from which material flows can be routed in a timely manner to appropriate recycling plants. To address these issues, analysis based on Big Data or RecycleGO is proposed to determine waste routing [15].

Social factors relate to motivating consumers to participate in separate collection. At present, they are not actively involved because of weak regulation [25], lack of customer awareness, an insufficient number of blockchain specialists, and protection of consumer data under the General Data Protection Regulation (GDPR) [24]. Citizens’ willingness to take part can be encouraged through token rewards and other digital payment mechanisms, or even by penalizing non-compliant users, similar to ride-sharing systems such as Uber (Lamichhane, Sadov, and Zaslavsky) [27]. Applications already exist that link citizens with waste collection companies, for example, SGB Simulation or an Ethereum private network using Geth within SGB itself, where sensors send data to a Raspberry Pi [28]. Nevertheless, with a suitable incentive program, proper collection, penalties for non-compliant users, rewards, and other measures, the model can become an “Uber for waste.” A new approach is a data-driven artificial intelligence tool for assessing people’s socio-economic attitudes and the compatibility of different systems [8].

A major factor is the financial one, as substantial funds are required to deploy and maintain blockchain systems. For this reason, the article proposes a decision algorithm for applying blockchain or a database.

Depending on the specifics of the flows being traced, the entities involved, and participants’ access rights, different platforms apply—with open, private, or hybrid access. The rule is that data are correctly recorded on the blockchain and then shared. In this respect, a decision is needed on whether a given DLT platform or database is appropriate [24]. The presence of a central authority does not by itself dictate the use of a database, because the communicating parties include consumers, processors, producers, and state and regulatory bodies, especially when tasks such as billing and documentation are involved [14], and there is a need to remove a third coordinating party between different actors in the system [3].

Incentives to adopt DLT regardless of cost include new regulations on extended producer responsibility for the life cycle of products they manufacture, encouraging customers to relinquish ownership by handing products over for recycling [24] and subsequent reward through token payments or other digital currencies [14]. Additional support comes from trends toward digitization and communication in waste collection and the development of smart cities [14,27,29], based on blockchain and IoT (Internet of Things Lamminchane et al.) [30].

This is why this article examines the reasons for choosing DLT and proposes a design aid algorithm for its application [27].

It can be stated that the decision process is governed by five main factors: processing time, cost, security, and reliability of the platform [3], plus an additional factor proposed here—complementary currency (proprietary currency) [23], including token payments in communication with consumers [14]. The process is shown in Figure 3.

Unresolved issues in this field include a shortage of trained specialists and the ability to integrate different technologies, including AI [9]. This means that building a network for monitoring the status of waste bins, routing, and assigning recycling plants could be managed by artificial intelligence. Constructing such a network requires not only skilled personnel but also an available and correctly populated database of plant capabilities, their processing parameters, timing, waste characteristics, materials, mixed products, and other data. The most important factor is compatibility among all software components (IoT, logistics, analysis, blockchain), a properly selected platform, participants’ rights, and above all, the right to choose the solution.

The connection between sustainable development and blockchain has been analyzed by Flourentzou in [31].

Table 1. Classified bibliography by field.

№	Reference	C	Tr	SD	R	CE
[1]	Khan, I.A.; Haq, F.; Kiran, M.; Aziz, T. Circular Economy and Waste Management: Transforming Waste into Resources for a Sustainable Future.	3	1	2	4	6
[2]	Sadeghi, M.; Mahmoudi, A.; Deng, X. Adopting Distributed Ledger Technology (DLT) for the Sustainable Construction Industry: Evaluating the Barriers Using Ordinal Priority Approach	1	1	6	1	2
[3]	Gopalakrishnan, P.; Hall, J.; Behdad, S. A Blockchain-Based Traceability System for Waste Management in Smart Cities	5	2	3	3	5
[4]	Europe 2020 Strategy. Available online: https://ec.europa.eu/info/ (accessed on 16 May 2026)	2	0	3	2	5
[5]	A French Act of Law against Waste and for a Circular Economy. European Circular Economy Stakeholder Platform	1	0	2	2	4
[6]	People’s Republic of China. Circular Economy Promotion Law, 2008	2	0	2	2	4
[7]	OECD Annual Report. Available online: https://www.oecd.org/newsroom/43125523.pdf (accessed on 16 May 2026)	1	0	2	2	4
[8]	Kulwant, M.; Raj, D.; Yadav, A.K. Waste Management and Recycling	2	0	2	2	4
[9]	Bułkowska, K.; Zielińska, M.; Bułkowski, M. Blockchain-Based Management of Recyclable Plastic Waste	2	1	0	2	0
[10]	Saboor, A. Sustainable Waste Management Using Blockchain Technology in ASEAN Nations: Case Study of Indonesia	2	0	4	2	4
[11]	Ugozor, A.O.; Udebuani, A.C.; Nnorom, O.O. Production of Bioplastic from Waste Agricultural Biomass	1	0	0	1	0
[12]	Liang, W.; et al. How to Track the Dumped Waste? From Data Collection to Advanced Source Identification	1	1	0	0	2
[13]	Khudyakova, T.; Lyaskovskaya, E. Improving the Sustainability of Regional Development in the Context of Waste Management	1	1	1	0	1
[14]	Berg, H.; Sebestyén, J.; Bendix, P.; Le Blevennec, K.; Vrancken, K. Digital Waste Management	6	1	1	1	1
[15]	Gong, Y.; Xie, S.; Arunachalam, D.; Duan, J.; Luo, J. Blockchain-Based Recycling and Its Impact on Recycling Performance: A Network Theory Perspective	4	3	0	3	0
[16]	Enache, B.-A.; Banica, C.-K.; Bogdan, A.G. Intelligent Trash Level Detection for Sustainable Urban Waste Management	1	0	1	2	0
[30]	Hansen, J. Brilliency Wallet to Include Swytch’s Blockchain-Based Token	1	1	1	2	0
[17]	Obiorah, C.A.; Ndubuisi, O.G.; Ali, S.E.; Aku, U.T.; Nesiama, O.; Agbakhamen, C.O.; Gumi, S.A. Integrating Waste Management into Sustainable Development Planning: A Framework for Urban Areas	1	0	1	2	0
[18]	Flynn, S. 3 Ways the Recycling Supply Chain Is Using Blockchain	2	1	0	1	0
[19]	German Association Using IOTA DLT to Incentivize Plastic Recycling	1	2	0	0	0
[20]	Preece, J.; Easton, J. To Blockchain or Not to Blockchain, These Are the Questions: A Structured Analysis of Blockchain Decision Schemes	2	0	0	2	1
[21]	Ntshangase, S.; Ndhlovu, N.; Myaka, S.; Mahlasela, O.; Siphambili, N.; Mthethwa, S. An Evaluation of DLT Governance Models	1	1	0	1	0
[22]	Wang, B.; Wang, N. Decision Models for a Dual-Recycling Channel Reverse Supply Chain with Consumer Strategic Behavior	1	1	0	1	0
[23]	Khadke, S.K.; et al. Efficient Plastic Recycling and Remolding Circular Economy Using the Technology of Trust—Blockchain	4	2	0	2	0
[24]	Taylor, P.; Steenmans, K.; Steenmans, I. Blockchain Technology for Sustainable Waste Management	5	2	0	2	0
[25]	Chidepatil, A.; Bindra, P.; Kulkarni, D.; Qazi, M.; Kshirsagar, M.; Sankaran, K. From Trash to Cash: How Blockchain and Multi-Sensor-Driven Artificial Intelligence Can Transform Circular Economy of Plastic Waste?	2	2	0	1	0
[26]	Zhou, Z.; et al. Environmental Performance Evolution of Municipal Solid Waste Management by Life Cycle Assessment in Hangzhou	1	1	0	1	0
[27]	Lamichhane, M. A Smart Waste Management System Using IoT and Blockchain Technology	3	2	0	1	0
[28]	Li, J.; Kassem, M.; Ciribini, A.L.C.; Bolpagni, M. A Proposed Approach Integrating DLT, BIM, IoT and Smart Contracts: Demonstration Using a Simulated Installation Task	2	1	0	0	0
[29]	Ahmad, R.W.; Salah, K.; Jayaraman, R.; Yaqoob, I.; Omar, M.A. Blockchain for Waste Management in Smart Cities: A Survey	1	1	0	1	0
[31]	Flourentzou, P. Decentralized Sustainability: Integrating Climate Action into Digital Asset Management on Distributed Ledger Technology (DLT)	2	3	1	2	0
[32]	Chowdhury, M.; Colman, A.; Kabir, M.; Han, J.; Sarda, P. Blockchain Versus Database: A Critical Analysis	1	0	0	0	0
[33]	Tijan, E.; Aksentijević, S.; Ivanić, K.; Jardas, M. Blockchain Technology Implementation in Logistics	2	1	0	0	0
[34]	Chu, T.; Ma, J.; Zhong, Y.; Sun, H.; Jia, W. Shared Recycling Model for Waste Electrical and Electronic Equipment Based on the Targeted Responsibility System in the Context of China	1	3	0	1	0
[35]	Riley, L. Universal DLT Interoperability Is Now a Practical Reality	1	1	0	0	0
[36]	Hristov, P. Data Management Model Based on Distributed Ledger Technologies	1	1	0	0	0
[37]	Fox, A.; Brewer, E. Harvest, Yield, and Scalable Tolerant Systems	2	1	0	0	0
[38]	Mulligan, C. These 11 Questions Will Help You Decide If Blockchain Is Right for Your Business	7	3	0	0	1
[39]	Peck, M. Do You Need a Blockchain?	5	1	0	0	0
[40]	Koens, T.; Poll, E. What Blockchain Alternative Do You Need?	2	2	0	0	0
[41]	Yaga, D.; Mell, P.; Roby, N.; Scarfone, K. Blockchain Technology Overview.	3	1	0	0	0
[42]	Lewis, A. The Basics of Bitcoins and Blockchains: An Introduction to Cryptocurrencies and the Technology that Powers Them	2	1	0	0	0
[43]	IBM Blockchain—Enterprise Blockchain Solutions & Services	2	1	0	0	0
[44]	Suichies, B. Why Blockchain Must Die in 2016	2	0	0	0	0
[45]	Pons, J. Blockchain and DLT Application to Sustainable Development: New Architectures and Use Cases for Circular Economy	3	0	2	4	4
[46]	Nakamoto, S. Bitcoin: A Peer-to-Peer Electronic Cash System	2	1	0	1	0
[47]	Parulava, S.; Brown, R.; Birch, D. Towards Ambient Accountability in Financial Services: Shared Ledgers	1	0	0	0	0
[48]	Castro, M.; Liskov, B. Practical Byzantine Fault Tolerance	2	1	0	0	0
[49]	Blockchain and the GDPR	2	2	0	0	0
[50]	Anceaume, E.; Del Pozzo, A.; Rieutord, T.; Tucci-Piergiovanni, S. On Finality in Blockchains	3	0	0	1	0
[51]	Androulaki, E.; et al. Hyperledger Fabric: A Distributed Operating System for Permissioned Blockchains	3	2	0	1	0
[52]	Garay, J.; Kiayias, A.; Leonardos, N. The Bitcoin Backbone Protocol: Analysis and Applications	1	0	0	1	0
[53]	Dwork, C.; Lynch, N.; Stockmeyer, L. Consensus in the Presence of Partial Synchrony	4	0	0	1	0
[54]	Malavolta, G.; Moreno-Sanchez, P.; Kate, A.; Maffei, M.; Ravi, S. Concurrency and Privacy with Payment-Channel Networks	1	0	0	0	0
[55]	Maiti, I.; Kotliarov, I.; Lipatnikov, V. A Future Triple Entry Accounting Framework Using Blockchain Technology	1	1	0	0	0
[56]	Interoperability of Blockchain Solutions. European Union Blockchain Observatory and Forum	1	1	0	0	0
[57]	Bellavista, P.; Corradi, A.; Foschini, L.; Monti, S. Improved Adaptation and Survivability via Dynamic Service Composition of Ubiquitous Computing Middleware	1	0	0	0	0
[58]	Nwufoh, P.; Hu, Z.; Wen, D.; Wang, M. Nanoparticle Assisted EOR during Sand-Pack Flooding: Electrical Tomography to Assess Flow Dynamics and Oil Recovery	1	0	0	0	0
[59]	Caldarelli, G. Understanding the Blockchain Oracle Problem: A Call for Action	1	1	0	0	0
[60]	Vogels, W. Eventually Consistent	1	0	0	0	0
[39]	Peck, M. Blockchain World: Do You Need a Blockchain? This Chart Will Tell You If the Technology Can Solve Your Problem	1	1	0	0	0
[61]	Kannengießer, N.; Lins, S.; Dehling, T.; Sunyaev, A. Mind the Gap: Trade-Offs between Distributed Ledger Technology Characteristics	1	1	0	0	0
[62]	Projects—Sun Exchange	1	1	0	0	0
[63]	The Capital Market for Every Asset on Earth—Solana	1	1	0	0	0
[64]	WHES—Intelligent C&I Energy Storage Manufacturer & Solution Provider	1	1	0	0	0
[65]	Stone, L. Energy Web Foundation Launches Pioneering Blockchain to Accelerate a Low-Carbon	1	1	0	0	0
[66]	Linnet, M.E.; Wagner, S.; Haswell, H. Maersk and IBM Introduce TradeLens Blockchain Shipping Solution	1	1	0	0	0
[67]	Flaherty, N. Volkswagen Uses Blockchain for Automotive Supply Chain	1	2	0	0	0
[68]	Blockchain in the Energy Sector—Real World Blockchain Use Cases	1	2	0	0	0
[69]	Švažas, M.; Makutenas, V. Energy Systems in Agriculture: Synergies and Perspectives	1	0	1	0	1
[70]	Bhubalan, K.; et al. Leveraging Blockchain Concepts as Watermarkers of Plastics for Sustainable Waste Management in Progressing Circular Economy	1	0	0	1	0
[71]	Sultana, A.; Barua, M. Emerging Technologies in Sustainable Finance: AI and DLT as Enablers of the Green Economy	1	1	0	1	1
[72]	Bodislav, D.A.; et al. Recyclable Consumption and Its Implications for Sustainable Development in the EU	1	0	0	1	1

classifies the literature by field, where C is citation, Tr is track, SD is sustainable development, R is recycling, and CE is circular economy. The table classifies the literature sources regarding the use of blockchain for tracking materials to achieve the criteria of sustainable development through the circular economy and recycling.

The categories in Table 1 (C, Tr, SD, R, CE) are classification indicators, not empirical measures. They are classified from the reviewed literature and cannot be interpreted as ratings of technological efficiency.

3. Results

The choice of a suitable platform depends on the application domain (waste management). Main objectives, such as transparency, trust among partners, elimination of a third party, etc., and others are incorporated as criteria in the decision support tree [23]. Reference [31] proposes a decision support tree for determining how information is transmitted and stored—databases or blockchain—the main criterion being the presence of multiple participating parties. For choosing network types—public, private, hybrid—the arguments are: trust between parties, trust in the third party, data immutability, scalability. The network type is determined according to the permanence of records. The linear structure omits criteria like robustness and fault tolerance, the presence of trust [29,32] cooperation, inertia to change, security, assurance of information quality, and the balance of risks and benefits. Therefore, decision models for various applications were reviewed, and a model was structured according to the specifics of waste collection. The decision support tree proposed in this article is aimed at achieving: (1) clarity in ownership rights over products and waste, (2) support for legal and policy goals by encouraging sustainable waste management, (3) preservation of anonymity and confidentiality for institutions and individuals [24,33], (4) facilitation of payments or rewards [15] okens [14], (5) waste monitoring and tracking (e.g., SNCF initiatives and the Dutch Ministry of Infrastructure.

The proposed algorithm is suitable for use by recycling companies, taking into account the characteristics of partners such as:

• Producers: Design for recycling, use of recyclates, and recycling of production waste.
• Consumers: Improving separation behavior and purchase decisions, which can be encouraged as part of a community that supports higher environmental standards and customer engagement [32].
• Recyclers: Improving the sourcing of suitable waste-streams [14]. We focus on producers because they record waste collected at the input stage and process it.

The stakeholders listed above can successfully form the basis of a communication network. When different types of waste are present—with varying consistency, origin, and treatment method—they must be routed correctly to plants in order to save costs and increase efficiency. The separation center is introduced as a dedicated life cycle actor responsible for batch qualification, routing, exception handling, and audit-ready registration (

Table 2. Role of the separate center.

Function	Input	Interaction with	Output or DLT Record
Batch intake	Delivery and digital batch ID	Transport company, producer/consumer records	Record for intake confirmation
Material qualification	Sensor or manual with QR, RFID, composition data	Monitoring entity, plant capability database	Material profile mi
Record validation	Batch metadata, origin, quantity, timestamps	Producers, consumers, regulators	Valid or invalid status
Routing decision	mi, plant capability, availability, regulation	Recycling plants of n, composting, landfill	Routing instruction according D(Bi)
Exception handling	Plant outage, overload, new technology availability	Transport operator, alternative plants	Re—routing event
Audit & reporting	Routing history, processing route	Regulators, third-party auditors	Immutable audit transaction
Incentive linkage	Verified collection	Payment or token	Reward (no PII)

, Figure 4). Its routing decision D(B_i) is defined in Equations (1) and (2) (in Figure 4). In this case, the proposed separation center is a central player in the life cycle because it is a node for routing, material type determination, logistics to a nearby and compatible processing technology, and data recording when evidence is needed. This participant should not be tied only to sorting or to the output of other activities in the recycling chain, as in [33]. The sorting entity collects all material flows from the producer, retailer, consumer, and the third monitoring party and assesses efficiency. In that model, the distribution function of the separation entity is missing; by identifying the material type, it can route it correctly to the right processing site. The reason is that waste contains a mix of materials, not all plants can process them, and this reduces their efficiency.

In this study, the proposed separation center is classified as a specialized life cycle actor. Its functions are to classify incoming batches according to material composition and to correctly direct each qualified batch to the possible processing stages. These are recycling, composting, landfilling, repackaging, or reuse plants.

If we denote the life cycle graph by G, define the network participants by N, and the data and material exchange by E, we obtain:

G = (N, E),

(1)

where N are producer, consumer, transport operator, recycling plant, regulator, and monitoring organization.

For incoming batches B_i at the center, the transport operators must return a routing decision D(B_i) = {destination, processing route, priority, audit trail}.

Thus, the center’s responsibilities are:

Qualification of materials according to type, level of contamination, and compatibility with available processing technologies.

Record validation—checking of incoming digital records

Routing—selection of an appropriate facility, which means recycling, composting, reuse, or referral for incineration.

Life cycle registration—creation of a traceable, immutable record in the DLT layer.

Exception handling and redirection—in case of unavailability or overload of the target plant.

Regulatory reporting—recording of data in a form so that evidence can be presented without sharing sensitive data.

Routing logic

The routing decision for each batch B_i is determined by:

D(B_i) = f(m_i, C_j, A_j, R_k, S)

(2)

where:

m_i = material type with quality attributes of a given batch i;

C_j = processing capacity allocated to suitable plant j;

A_j = status of plant j for acceptance, operation, interruption, capacity;

R_k = regulatory and environmental constraints;

S = confidentiality requirements for batch data.

The routing procedure is given by the logic If–Then:

If the batch record is incomplete or unverifiable, then rejection and a request for correction.

If the material type is incompatible with all active plants, then routing to composting, landfill or temporary storage.

If more than one compatible plant exists, then the one with the lowest transportation costs is selected.

If the selected plant turns out to be unavailable, then redirect to the next suitable plant.

Receive feedback on the arrival of the load sent for processing.

With the roles thus described, the need for deterministic transaction finality is determined, and in many cases, a multichannel DLT architecture is allowed (Section 4.3). Voluntary and paid citizen participation is a complementary asynchronous layer (Figure 5). The role of the separation center is given in Table 2.

4. Discussion

For the correct choice of solution when implementing blockchain at the input of processing facilities, some features of the chain were included: collection, separation, routing to suitable processing plants through digital technologies, recycling, landfilling, or repackaging into new products and raw materials. Because of the diverse stakeholders, a simplified decision algorithm is required.

4.1. DLT—Selection Criteria

In well-known algorithms, the leading criterion is transaction speed, but in the case proposed, trust among unknown and newly joined network participants is of utmost importance. This is achieved by controlling data confidentiality and data immutability. For simplification, a circular model is proposed in which every “YES” answer recommends a type of DLT. The criteria in this case are as follows:

-

Transaction traceability—The desired traceability of transactions recorded on the chain is achieved through DLT that enables trusted traceability among participating parties.

-

No interoperability—There is no operational interoperability because the systems are new. Today, DLT relies on various methods to achieve interoperability [34,35].

-

CAP Theorem—Consistency—Every read receives the most recent write or an error. Availability—Every request receives a (non-error) response, without the guarantee that it contains the most recent write. Partition tolerance—The system continues to operate despite an arbitrary number of messages being dropped (or delayed) by the network between nodes [36].

-

Transaction speed—DLT technologies are not suitable for managing processes with rapidly changing parameters. Data often change at high speed (>100 ms), which is an additional criterion. DLT transactions are slower, but this speed is appropriate for the material flows herein considered [37,38,39].

-

Data volume—In the circular economy system, specific pre-selected data types by category are used, whose volume grows only up to certain limits [37].

-

Immutable data—There is a requirement for data immutability. This means that data entered, verified, and stored cannot be modified and are governed by network rules. New data may be added, but previously entered data remain unchanged [37,40].

-

Data type (single-type/multi-type) [37] The homogeneous data structure is among the drawbacks of DLT technologies because different DLT archetypes exist and their processes differ significantly. Work is underway on a unified archetype to provide greater flexibility, scalability, and transaction processing capacity.

-

No PII (personally identifiable information)—Under the EU’s GDPR (and similarly in the United States), information that leads to personal data must not be stored on DLT because it is immutable.

-

Master–master data insert—Each participant can insert data using an identification key. For digital collection rates, this is an important facilitation for consumers. Every operator with a digital key can enter data, while other participants verify them for inclusion on the blockchain [40,41]

-

Confidentiality requirement—In the case proposed this is very important for both consumers and recycling enterprise [38,42]

-

Need to retain history and a transaction register—These records are real accounts. They are used to obtain tokens, rewards, penalties, reports to regulators, and so on [40]—Model DHS.

-

Need to remove intermediaries—This leads to increased trust among network participants [37]—Cathy Mulligan.

Analysis with VOSviewer (version 1.6.21) shows the parameters with the greatest weight. They are the CAP theorem, immutable data, traceability, no interoperability, maturity, and history retention; these are followed by: PII, speed of transaction, etc. The most important are shown in Figure 5. They are evidence of the importance of the selected criteria and will not be used for validation or causal relationships in subsequent analyses.

4.2. Scheme

Figure 6 visualizes the decision-making scheme. If the answer is Yes, enter the quadrangle and select a DLT application. If the answer is NO, a database is selected.

After the digital solution is determined—databases or DLT—a suitable platform must be chosen: Blockchain, Ethereum, IOTA, or others. The common factors among DLT platform types are cost—for removing intermediaries, paying for new products [43] and energy consumption [44]. With rising electricity prices, this parameter is one of the most important. In the proposed linear decisionsupport tree for choosing public, private, hybrid DLT, or another option, the criteria are as follows:

-

Can everyone use the data [45]? How many people use the data [37,38]?

-

How many people have the right to modify the data?

-

Right to admit a new member—intrafirm or interfirm organization, or control over business logic; we propose this node as the option to include new suppliers.

-

Cost—is it energy-intensive or not (IBM) and return on investment after maintenance cost, new equipment, and benefits are defined: will the system justify the price [38,41]? Security data—must the data be secure (linked to security procedures [38])?

4.3. Applicability of Criteria for Recycling

In Section 4.3. the theoretical classes of synchronicity, properties of finality, and specific DLT architectures are distinguished. Since the choice of network depends on the implementation and not on universal rules, there is no categorical definition of a network type.

This section of the article examines the Hyperledger Fabric platform from the perspective of the three main layers of analysis, such as synchronicity in the distributed ledger [46], transaction finality features [47,49], and implementation of DLT according to permissioned and permissionless networks.

The applicability of the model is determined by the network type, which requires in-depth analysis. One of the most important criteria is data immutability, which is impossible without a new transaction that is in fact, hashed. In a synchronous chain, after a request is finalized, data remain append-only within the model, and if reality is violated, the boundaries Δ are lost. This means that the chain is in conflict with the right to erasure [46]. In a semi-synchronous chain, periods before GST (Global Stabilization Time) stand out, with the possibility of block reversal, but after that blocks are fully irreversible [47]. In asynchronous chains, before the longest branch is recognized, blocks may be dropped, but after that, recognized and recorded blocks remain [48,49]. With respect to product life cycle data, this means that popular asynchronous networks are not always the most suitable. For the recycling enterprise and the separator, it is particularly important that data remain recorded and that queries can be made in subsequent audits. The separator is the main routing entity, after the transport company, directing flows to the correct plant.

Homogeneity is supported by block validity—that is, format, state, and validity rules—but in heterogeneous systems, different channels are provided for data transfer. Public single-ledger models are best suited for sharing manufacturing data, especially in the recycling industry. Hyperledger Fabric features channel separation, private data collections, and differentiated approval policies, allowing for the privacy of stakeholders in a consortium [50,51]. Hyperledger Fabric is an enterprise-grade permissioned consortium suitable for auditing and routing solutions. It is not a strictly synchronous network.

In synchronous networks, heterogeneity is not expected; blocks look the same within a single round, and the approach is application selection. In partially synchronous networks, the approach is again application selection, but after GST, everyone shares the same ordered SMR log. In asynchronous networks, all blocks converge to a canonical structure; the protocol formats a single view depending on the application, and transmission depends on the protocol and implementation. Given the presence of different materials, synchronous and asynchronous networks would not satisfy the conditions of the environment. It goes without saying that new materials are continually being developed, requiring modern recycling and repackaging technologies, and the possibility of including new participants.

Speed can be regarded as a function directly proportional to throughput and recognition of completion—that is, transaction finality [46,49,52]. The main parameter defining this factor is hardware load, which depends directly on the consensus protocols applied. It is high in synchronous networks but depends on the bound Δ and other factors. In partially synchronous networks, it is high after stabilization, high for permissioned benchmarks, and moderately high predictability after stabilization. In asynchronous networks, uncertainty ranges from seconds to days depending on the application. Block size can be configured in the same way, while predictability is low under a strict upper bound. For example, when assigning a recycling plant, high speed and transaction finality are required; payment to citizens may be deferred for a short period. The presence of many waste collection companies and citizens suggests building an asynchronous part of the overall network.

The CAP theorem shows that linear consistency (C), availability (A), and partition tolerance (P) cannot all be satisfied at once; the system chooses behavior according to the implementation [36]. This is the rule for system survival when part of it is disconnected and later resynchronizes when a node that was offline in time reconnects. In synchronous networks, when no partition occurs, analysis is possible, and partitioned nodes are ignored. Consistency and tolerance to delay are also characteristic of partially synchronous networks after GST; before that, unpredictability is typical. When building synchronous and partially synchronous enterprise networks, C + P is often used. In asynchronous networks, consistency is impossible without assumptions, especially during system failure or when a shorter chain is dropped. This means that when tracking waste in transport from point A to plant B or C, an asynchronous chain may share incorrect data or data from a previous record. For example, at point A, plastic waste is handed over to plant B, but at the same time, plant B has an outage and cannot accept the delivery intended for it. A new plant, plant C, implements a modern technology for processing the same type of plastic waste. When data are shared, the link among the three entities is lost—that is, trucks do not receive a message from plant B about the outage. Because that part of the chain is dropped (data are not shared and availability is missing), the material is delivered to plant B rather than to plant C, which has the processing technology. In an asynchronous network, elements that cannot synchronize and update their information must be stopped. This means sacrificing availability in favor of consistency. This example illustrates that the trade-offs between consistency and availability are a risk to timely routing updates in the absence of off-chain reconciliation or coordination mechanisms. Each example depends on the application level.

Consistency and finality of networks are considered for synchronous, asynchronous, and partially synchronous networks [49]. In BFT-based partially synchronous consortium networks, blocks can achieve deterministic finality once consensus is reached and the ordering service has stabilized after GST. In contrast, Nakamoto-style Proof of Work networks provide probabilistic finality because forks and chain reorganizations remain possible until sufficient confirmations are accumulated [49]. In partially synchronous networks (permissioned, e.g., Red Belly, Hyperledger Fabric), deterministic immediate consistency is achieved. In open networks with high energy consumption, priority is given to availability/progress and forks are allowed—that is, consistency may be weakened, and blocks may be reverted.

Performance is examined from different aspects depending on the network type [50,52]. These include packet delay, predictability, packets, and Δ-time. Throughout this section, GST denotes Global Stabilization Time, and Δ denotes the upper bound on message delay after GST in the partially synchronous model [47]. The main characteristics are summarized in

Table 3. Network timing and packet behavior (English).

Aspect	Synchronous Network	Partially Synchronous Network	Asynchronous Network
Packet delay guarantee	Every message arrives within ≤Δ (known Δ)	Delay is bounded, but after GST, there is an unknown Δ	No upper bound; a packet may be delayed arbitrarily; with a reliable channel, delivery still occurs
Transmission predictability	High: known maximum delivery time	Medium: unpredictable before GST, stable after GST	Low; in many models, loss occurs
Packets in a blockchain context	Blocks, votes, transactions—fixed timeouts used	Blocks, votes, transactions—adaptive timeouts; stabilization via BFT rounds with expected latency	Many timeouts, many retransmissions; the rule does not apply; no “single Δ for all”
Link to data finality	Planned; time for commit: “after k·Δ all honest nodes have seen the block”	After GST, deterministic finality under BFT	Probabilistic finality: “final” after N confirmations on the longest chain
Risk of reordering/late packets	Limited within Δ	Limited after GST; high before GST	High, arrival of an old block leads to chain re-evaluation (Nakamoto)

Legends: GST = Global Stabilization Time; Δ = delay bound (define once in the paper); BFT = Byzantine fault tolerance.

.

The three types of blockchain networks can also be compared on the interoperability criterion, which is very important for the future construction of an operational network that receives data from sensors [52,53,54,55]. This property is linked to the real ability to manage and configure parameters and to achieve higher efficiency, as shown in

Table 4. Criteria for a real ability for efficiency.

Criterion: Interoperability	Synchronous Network (Known Δ)	Asynchronous Network (No Delay Bound)	Partially Synchronous Network
Cross-chain coordination	Common timeouts and rounds are easy to set through synchronized clocks between parties.	There is no reliable message bound; reliance is on cryptographic proofs, not on “everyone has seen by time T.”	After stabilization, BFT is easier to plan; before GST, behavior is unpredictable.
HTLC/time locks	With a realistic Δ, time windows are aligned.	Risk of late messages when timeouts expire; time-jacking attacks, because there is no unified clock.	Compromise: after GST, delays are more stable in consortium networks.
Notary/relay/light client	In controlled infrastructure (data center, private links), there are few problems.	On the public internet, proofs of inclusion, finality, etc., are required.	Predictable delays after stabilization.
Standards and layers (ISO, W3C, common formats)	Not directly dependent on Δ, but applied in test networks with fixed timeouts.	Focus on formats and semantics; time is off-network uncertainty.	Enterprise standards and operational SLAs after stabilization.

.

Operational and legal traceability is achieved in a closed consortium under a partially synchronous model with BFT [56,57]. Open networks with IoT approximate an asynchronous model, with traceability provided by cryptographic trails and consistency among devices, while strict synchrony applies only locally or under theoretical assumptions.

Synchronous network: Suitable for analysis in an isolated environment (internal network with a single provider), but unrealistic at a global scale.

Partially synchronous network: Closest to industrial traceability platforms (permissioned DLT, BFT ordering), which makes them suitable for waste processing across multiple stakeholders. Such a network offers predictable recording time and easier consortium audit, and it is characterized as “stabilizing.”

Asynchronous network: Close to the real environment; traceability is ensured by immutable records + signatures + sometimes anchors on a public chain, not by “everyone synchronized to the second.” Eventual confirmation relies on proofs, not on a clock.

The presence of a third party that distributes blocks often causes delay, whereas the absence of centralized control makes deliveries slower [59]. In the case of managing a chain with multiple inputs, an oracle is required. In a synchronous network, messages arrive within a known time when intermediaries exist, under a realistic Δ. Unfortunately, strict synchrony on the open internet is unrealistic, which makes these networks virtually impractical.

In a partially synchronous network, stabilization is based on BFT protocols. The advantages compared to “intermediary-free” systems are:

• There are zero intermediaries in one sense, but also a defined set of validators with a special role in organizing sharing with distributed trust—not literally “zero intermediaries” [5].
• For firms, it is enterprise-grade, which suits audit and SLA requirements.

In an asynchronous network, there is no upper bound on delay, deterministic consensus is impossible, and the advantages are the absence of a central transaction orderer, with continuation and propagation of the transaction from the longest chain when there is no intermediary. In no way does this remove the need for intermediaries and oracles outside the chain.

The factor discussed is especially important for the proposed system because payment to waste collection companies and citizens is possible through an application with an asynchronous system, but the parameters of recycled raw materials require strict accountability through a partially synchronous network. Another issue is routing output products to incineration or repackaging, which sometimes requires confidentiality and data sensitivity. It may even be necessary to build part of the system as a synchronous network with a trusted data type.

Master–master insert is one of the most important criteria because it determines the order of records while following a conflict policy—whether based on CRDTs, time, versions, or other approaches [60]. When an upper bound on delay is required, the degree of network synchrony is determined accordingly.

In synchronous networks, everyone has seen everything, and a unified insert order is easier to achieve. The problem in a real network is the unrealistic assumption of exchange and transaction submission without satisfying synchrony. They are suitable for small closed networks. The partially synchronous model introduces delay, but through BFT or state machine replication protocols, a consensus order is applicable, which makes it the most common model for enterprise master–master behavior on a shared registry [47,52]. Asynchronous chains do not provide order by themselves; order must additionally be implemented and defined cryptographically.

Personally identifiable information (PII) cannot be classified by network type; it is governed by the GDPR [48,50]. In synchronous networks, without a real internet connection, deletion or minimization of data is possible, whereas in partially synchronous networks, the consortium is the administrator and controls the channels [50], with roles and participants having restricted access. In asynchronous networks, it is difficult to guarantee the identifiability of participants. GDPR compliance of DLT is not an automatic property, but depends on the design, roles, volume, and routing of data access, etc. For improved compliance, data minimization, separation of integrity evidence from personal data, access roles, or presence of an oracle, clear responsibilities are recommended.

The last criterion is data volume, but this is rather an economic parameter, although it is known that small packets are transmitted reliably and securely, unlike large ones, and do not form queues. Therefore, when sharing data, regardless of the network type, smaller data blocks should be prepared.

Due to their nature, data flows for the separation hub require a permissioned and partially synchronous DLT environment with deterministic finality. The reasons range from the possibility of requesting a downstream audit to proof of shipments by the listed factors—type, processing technology capacity, status, data sensitivity, and environmental and regulatory frameworks.

With this, the main proposed factors for choosing DLT or another solution outside this technology have been analyzed. The choice of a suitable solution depends on participants in the chain, data confidentiality, the amount of data, the need for transaction finality, the regulatory framework, and other factors. Nevertheless, the review presented here can support informed network design.

4.4. A Decision Support Tree

Section 4.2 establishes the database–DLT branch (Figure 6). This subsection extends the framework to network deployment types (Figure 7). To propose a new decision support tree, various algorithms for DLT selection were used, such as the models of Cathy Mulligan, IBM, Morgen Peck, Bart Suichies, DHS, Tommy Koens, and Erik Poll [37,39,42,43]. Positive aspects compatible with the raw material traceability procedure were adopted from each model.

A recent study on the nature and necessity of blockchain states that a balanced and transparent framework correctly guides the choice—or the decision to avoid it [20]. In the large chaos of requirements and options, it is important to select the right criteria which are grounded in the specific technology. Therefore, the aim of this study is to present major criteria for choosing blockchain rather than a database, based on the specifics of the raw material recycling process for building a circular economy.

Unfortunately, in this field, there has been little research in recent years on well-known models for DLT selection. Usually, the main components are compared (flexibility, opaqueness, performance, policy, practicality, security) [61], but it is found that not all factors can score well at the same time. One factor’s deterioration leads to the improvement of another. That is why the task is to create a simple decision support tree for choosing a suitable DLT for the needs of a multi-step, multi-factor process. It should be applicable to trained and untrained specialists alike, as well as to interested third parties.

Not all DLT criteria can be applied in a single decision tree. For example, in Cathy Mulligan’s model, it is stressed that records must be permanent, but not all records can remain permanent and immutable for regulatory, organizational, ethical, and other reasons. At a minimum, different countries apply different limits on harmful emissions. In addition, standards for water pollution, standards and technologies for waste processing, control measures, and other elements change continuously. The paper does not discuss changes in technologies and the invention of new devices that enable faster and cheaper recycling of already known materials. Another aspect is the invention of new materials produced either from old recycled stock or through new technologies. Therefore, the need to change standards, chains, and criteria must be considered when choosing a suitable DLT.

For the preparation and synthesis of the decision support tree, the first published government frameworks in the field were used. They do not address the problems of multi-stage production, which may not cover all the recycling steps described in the article. They also do not take into account changes in the regulatory framework, materials, and production technologies in the circular economy system to comply with the criteria for sustainable development. This leads to the need to synthesize a new solution based on the advantages of the known ones. Immutability is excluded from Mulligan’s model because it may not exist. The possibility of sharing is preserved. According to Peck’s and Koens & Poll’s model, the need for use and the exact architecture are selected. Given the future need for corporate process management in most recycling industries, the models of IBM, DHS, and enterprise-oriented frameworks are taken. The goal is for the network to be applicable, not idealized.

The novelty of the proposed tree is the following:

-

Its applicability in multi-stage and multi-cycle production;

-

Usability in hybrid DLT architectures with different data access rights;

-

A practical tool for selecting an appropriate architecture for managing data sharing.

A decision support tree is a method for rapid decision-making on whether to use a technology or whether it needs to be changed. It is very easy first to check whether, in a given case, a certain type of blockchain is appropriate or whether the system can be modified. It should be borne in mind that altering or implementing an entirely new system is not always possible. In the proposed new decision tree, basic questions are embedded. First among these is the requirement for full centralized control, which is unacceptable according to the theory of data sharing in a network. When several participants can record data, some form of blockchain is required.

The choice of a closed or open (permissionless) network depends on security rights. In a closed network there is a limited number of participants who share data among themselves at high speed, with high trust, and at low energy consumption. In open networks, data are shared more slowly and many participants can view them. The procedure for recognizing records across many nodes is associated with high energy consumption, and data reliability is diluted among participants. At the same time, when necessary, records can always be extracted and their existence proven. In open networks, there are two main algorithms: Proof of Work (PoW), with higher consumption, and Proof of Stake (e.g., Ethereum after the Merge), with much lower consumption. PoW is used by Bitcoin, Dogecoin, Litecoin, Bitcoin Cash (lower than Bitcoin, but still high), Monero, and others. They can be compared with closed (permissioned) networks with a high level of trust, such as:

• Hyperledger Fabric—a private multichannel network for business consortia;
• R3 Corda—with high trust among validators and strong identity control for financial institutions;
• Enterprise Ethereum—with restricted participants;
• Quorum—for corporate deployments;
• Multichain—public or private, for role and permission management.

In closed networks, data are confidential; in open networks, they are public.

It can be noted that most operational projects for energy sales and raw material traceability are built on open networks with free access, e.g., Sun Exchange (Bitcoin) [62], Solana [63], etc. The platforms include training, an academy, course verification, real market participation with various products (wallets, tokens, digital assets), and more. This represents a new level of communication between sellers and consumers.

Hybrid network types with increased trust among participants while preserving data include WePower (Ethereum) [64] and Energy Web Foundation (EWF) [65]. The highest-trust networks are of the closed type, such as ConsenSys Energy & Sustainability, used to trace Volkswagen parts or Maersk pharmaceuticals via IBM [66,67,68].

Hybrid systems are suitable for operating with parts and tracing raw materials. In these systems, sensitive data is managed by the closed part. Such data include SOC, cycles, alarms, power, contracts, tariffs, and aggregators built on Hyperledger Fabric or Quorum/Besu (permissioned). Management data cover types of materials for recycling or separation, number of processing cycles, use in new products, price, responsible persons, and other fields.

The open part is used for communication, transparency, traceability, and higher trust through publicly verifiable records. Anyone can verify hashes of reports on delivered/stored energy or materials, certificates of origin (REC/GO), and audit evidence for third parties. In these networks, an Energy Management System collects data for a defined period, records it on a permissioned ledger, and periodically publishes a Merkle root/hash to a public chain, which is provable on audit. This communication meets the conditions for transfer of recyclable materials because it provides in Figure 7:

• Reliable, traceable exchange among several organizations—plant, transport company, waste collection, recycling, reuse of materials, their incorporation into other products, and so on;
• Clear, available reports to regulators;
• Secure financial reporting;
• No disclosure of sensitive data.

It is important for users, manufacturers, and owners to navigate to the right technology easily and without any major problems. Therefore, a decision support tree suitable for multi-stage production is proposed for choosing the right framework for increased clarity, auditability, and energy efficiency. Figure 6 is the main separator between a database and a DLT, while Figure 7 serves to define the logic for implementing DLT, such as permissioned, public, or hybrid deployment.

For clarity of the decision-making logic, the explanation is introduced that technological and infrastructural barriers are responsible for the feasibility of the architecture. According to this, a database or type of DLT is chosen. Regulatory barriers determine access rights and confidentiality, which is partly also related to the perceptions of users, i.e., social barriers. The cost of development, implementation, maintenance, and energy efficiency of the architecture is the basis of financing, or these are financial constraints. Thus, according to the barriers introduced at the beginning of the article, the logic for making a decision on a suitable architecture is outlined. A database is preferred in cases of centralized management and a lack of requirement for multilateral shared recording with audit independence. DLT justifies itself in cases of multiple participants, the need for traceability, and a shared ledger without full central control.

Different scenarios are possible under the model, such as composting in small municipalities, large urban recycling of multiple materials, and a cross-border e-waste consortium. One could even try applying the model to replace diesel tractors with electric ones to reduce pollution in sectors such as agriculture, mining, smart cities, and others [69]. This can show the logic of the flows and the boundary conditions, but this is not an empirical validation.

The measurable indicators are in % for material recovery, correct routing, waste to landfill, completeness of traceability over a life cycle, data verification error, and citizen participation. Response time to an audit can be measured in days, and kilometers are measured by the distance traveled by cars.

The final questions and examples have demonstrated the viability of the proposed decision support tree for easy selection of the type of blockchain network depending on the application. The case considered is the overall framework for transparent, trusted, and verifiable tracking of materials from production through collection, separation, recycling, and subsequent use in products. Through DLT, the chain of many plastics can thus be traced unambiguously through more than 20 cycles, especially when they are labeled as monomer units [70]. This further shows that DLT is not only a new technology but also a means of saving raw materials and energy for future human welfare.

The proposed decision support tree is best taken as a structured path for choosing a digital solution in the field of materials recycling. This is not a rule for the best choice, because it depends on the technology and the specifics of the environment. In the presence of a separator, a further step is provided for data collection and validation of the algorithm.

In recent years, with the rise in AI, it has become possible to build a better-performing and more efficient network using generative algorithms, deep learning, machine learning, cloud technologies, and others. They can successfully support the management of the circular economy process [71]. When using these technologies, it is important to determine energy consumption as well and to define algorithms for circular recycling of energy and batteries [72].

Application examples clearly suggest the advantages of Hyperledger Fabric according to the criteria analyzed in Section 4.3 and the decision-making path shown in Figure 7. Data sharing in the field of recycling is possible in the absence of full central control and the presence of many participants, which excludes purely centralized databases. The presence of heterogeneous activities such as auditing, separation, partitioning, and others implies the requirement for multiple channels, which requires a permissioned, multichannel architecture with deterministic finality after stabilization. Such a consortium platform is Hyperledger Fabric. Through the asynchronous or public anchoring layer in Figure 7, the recommendations for requiring public verification, payment interfaces, or Merkle roots can be implemented. Thus, it is proven that the criteria analyzed in Section 4.3 and Section 4.4 is a decision-making path.

Limitations. The work presented is conceptual and has not been implemented, but the boundaries and recommendations set are a prerequisite for upgrading existing installations or building new ones. The transferability of the model depends on the regulatory framework, infrastructure maturity, and stakeholder requirements.

5. Conclusions

The presumption of care for future generations and preservation of resources underpins the management and recording of data from material recycling. The article proposes a conceptual procedure for choosing whether to adopt DLT or a database. A new, fast decision support tree is then proposed for selecting a suitable application in the context of implementing a circular economy according to sustainable development criteria.

Applying the procedure described is a first step toward choosing a database or a distributed data sharing network. This depends on the defined barriers to deployment and the aims of collection, recycling, and repackaging processes. Objectives may include logistics of waste from bins to landfill, traceability of waste during recycling, or end-to-end traceability of raw materials. The latter also requires citizen participation, which can be encouraged through token-based payment. It is a challenge to build a common system for realizing a circular economy in which all participants can share data and said data are visible, while sensitive metadata must be managed by a trusted circle of participants.

Difficulties in applying the circular economy for collection, recycling, and processing enterprises are defined. To achieve efficiency, digitization of collection, separation, and delivery to recycling plants is needed. Based on classified characteristics of the application domain, new criteria were developed for choosing databases or DLT.

A waste sorting center is proposed for the purpose of routing and allocation efficiency. The sorting center is an operational decision point because it connects material flows and their data to achieve digital traceability in the circular economy model.

Citizens can be motivated by providing rewards for the work they perform. These may be tokens or other incentives.

The presented framework for tracking and routing waste provides improved digital traceability to support the achievement of sustainable development criteria through digital identification. Unfortunately, primary data from processing is lacking, and currently there are no separation centers built in waste processing and landfill plants. With the proposed mechanisms and the introduction of an application for citizens to pay, we hope to stimulate their participation. The indicators provided in Section 3 and Section 4 are a starting point for assessing the impact of building a separation center.

The proposed model has measurable indicators of sustainability loss, such as misrouting rate, traceability, response time, etc. It is not complete, and the measured values also depend on other units in the scheme—landfills, transport, and external deliveries.

The approach to classifying criteria is by network type: synchronous, partially synchronous, and asynchronous. A detailed analysis was carried out of the three classified network types against the proposed criteria in the context of the problem considered. The choice is always justified when a multichannel data structure is present, as is typical for waste. One suitable platform in this case is Hyperledger Fabric, which enables simultaneous data sharing in a multichannel system among participants. Evidence of efficiency will be presented in a follow-up study.

On this basis, a superior conceptual decision support tree is proposed for choosing a database or DLT platform. A feature of the approach is that the system evolves, as do the platforms, so a single fixed solution cannot be prescribed. Unlike known decision trees, the proposed one is suitable for multi-step operations typical of material recycling while maintaining data transfer balance, auditing, and energy efficiency.

The developed model can be used both at the output and at the input of recycling enterprises. The suggested models are not claimed to be final; at the outset, and if necessary, other parameters may be included. The proposed framework offers a structured, recycling-specific path from barrier identification to architecture selection via combining routing logic at the separation center with DLT selection criteria under regulatory and data privacy constraints. Empirical validation with pilot routing and traceability data is planned as future work.

The next step is going to be a selection of the right BC platforms, routes, trained specialists, compatible software products, and greater citizen accountability.