Future Requests Research Articles

Dispatching AGVs with battery constraints using deep reinforcement learning

This paper considers the problem of real-time dispatching of a fleet of automated guided vehicles (AGVs) with battery constraints. AGVs must be immediately assigned to transport requests, which arrive randomly. In addition, the AGVs must be repositioned and recharged, awaiting future transport requests. Each transport request has a soft time window with late delivery incurring a tardiness cost. This research aims to minimize the total costs, consisting of tardiness costs of transport requests and travel costs of AGVs. We extend the existing literature by making a distinction between parking and charging nodes, where AGVs wait idle for incoming transporting requests and satisfy their charging needs, respectively. Also, we formulate this online decision-making problem as a Markov decision process and propose a solution approach based on deep reinforcement learning. To assess the quality of the proposed approach, we compare it with the optimal solution of a mixed-integer linear programming model that assumes full knowledge of transport requests in hindsight and hence serves as a lower-bound on the costs. We also compare our solution with a heuristic policy used in practice. We assess the performance of the proposed solutions in an industry case study using real-world data.

All-On-4 Implants Supported Prosthesis with a Modified Approach: A Case Report of Upper Jaw Rehabilitation

Young patients tend to prefer quick treatments and fixed restorations compared to older patients. Fixed reconstructions on implants now offer a variety of treatment options for various situations. However, despite the aesthetic and functional benefits, there are currently no clear clinical guidelines regarding the number of implants, loading type, and reconstruction type. Although various recommendations do exist, micro/macroscopic factors significantly influence treatment plans and surgical approaches in nearly every case. In this case report, we describe the rehabilitation of a triangular-shaped upper jaw of a 35-year-old non-edentulous patient using an implant-supported CAD/CAM designed prosthesis. Instead of using the All-on-4 method, five implants were inserted and immediately loaded after a series of extractions. The treatment plan involved placing three implants in the frontal region, with two tilted distally. The plan was designed not only to meet the patient's immediate demands but also to consider any possible future requests in the lower jaw. We also took into account the patient's current dentition and potential risk factors. As a result, we inserted a fifth security implant in the front region between the two front implants. The overall functional, aesthetic, and phonetic outcomes were satisfactory, and the patient expressed excellent satisfaction with the results.

RuleDRL: Reliability-Aware SFC Provisioning With Bounded Approximations in Dynamic Environments

As a key enabling technology for 5G, network function virtualization abstracts services into software-based service function chains (SFCs), facilitating mission-critical services with high-reliability requirements. However, it is challenging to cost-effectively provide reliable SFCs in dynamic environments due to delayed rewards caused by future SFC requests, limited infrastructure resources, and heterogeneity in hardware and software reliability. Although deep reinforcement learning (DRL) can effectively capture delayed rewards in dynamic environments, its trial-and-error exploration in a vast solution space with massive infeasible solutions may lead to frequent constraint violations and traps in poor local optima. To address these challenges, we propose a RuleDRL algorithm that combines the capability of DRL to capture delayed rewards and the strength of rule-based schemes to explore high-quality solutions without violating constraints. Specifically, we first formulate the reliable SFC provision problem as an integer nonlinear programming problem, which is proven to be NP-hard. Then, we jointly design DRL and rule-based schemes that are coupled to make the final decision and establish a bounded approximation ratio in general cases. Extensive trace-driven simulations show that RuleDRL can save the total cost by up to 65.67% and improve the SFC acceptance ratio by up to 82%, compared to the state-of-the-art solution.

OA-Cache: Oracle Approximation-Based Cache Replacement at the Network Edge

With the explosive increase in mobile data traffic generated by various application services like video-on-demand and stringent quality of experience requirements of users, mobile edge caching is a promising paradigm to reduce delivery latency and network congestions by serving content requests locally. However, how to conduct cache replacement when the cache is full is a challenging issue when faced with enormous content volume and limited cache capacity at the network edge while the future request pattern is unknown ahead. In this paper, we propose a cache replacement algorithm based on the oracle approximation named OA-Cache in an end-to-end manner to maximize the cache hit rate. Specifically, we construct a complex model that uses a temporal convolutional network to capture the long and short dependencies between content requests. Then, an attention mechanism is adopted to find out the correlations between requests in the sliding window and cached contents. Instead of training a policy to mimic Belady that evicts the content with the longest reuse distance, we cast the learning task into a classification model to distinguish unpopular contents from popular ones. Finally, we apply the knowledge distillation approach to assist in transferring knowledge from a large pre-trained complex network to a lightweight network to readily accommodate to the network edge scenario. To validate the effectiveness of OA-Cache, we conduct extensive experiments on real-world datasets. The evaluation results demonstrate that OA-Cache can achieve better performance compared to candidate algorithms.

Online Predictions for Online TSP on the Line

With a rapidly growing attention on the development of learning-augmented algorithms, we revisit the classical online traveling salesman problem (OLTSP) from the perspective of such learning approaches. A learning-augmented online algorithm, or simply an online algorithm with predictions, considers how to improve its competitive performance with theoretical guarantees by forecasting the information of future requests, e.g., their release time or locations in the OLTSP. In this study, we investigate the OLTSP on the real line, motivated by the parcel picking problem in a smart warehouse, which aims at scheduling a route for a server (saying, a Kiva robot), starting at the origin, serving all online requests and returning to the origin such that the completion time is minimized. Each online request is released over time on the real line and the server moves back and forth along the line with unit speed. We mainly focus on zealous algorithms, a special type of online algorithms for the OLTSP which never allow the sever to wait if there are still unserved requests, and exploit a specific forecasting strategy, called online predictions, which makes a sequence of predictions one by one in an online manner. In order to ensure better competitive performance, we particularly explore the worst-case scenarios that restrict the power of such predictions. Based on the discussion, we make an assumption in which online predictions are guaranteed to be useful, and devise a learning-augmented algorithm with [Formula: see text]-robustness and [Formula: see text]-consistency, [Formula: see text], comparing to the previous lower bound of [Formula: see text], where [Formula: see text] and [Formula: see text] which indicate prediction errors are sufficiently small constants. Moreover, we also conduct numerical experiments to demonstrate the practical effectiveness of the proposed algorithm.

Dynamic IoT service placement based on shared parallel architecture in fog-cloud computing

Fog-cloud computing is a promising computing paradigm for processing and storing massive data produced by Internet of Things (IoT) devices. Considering that the fog computing nodes are distributed, heterogeneous and have limited resources, it is not possible to deploy all the services belonging to IoT applications on the fog environment. Therefore, an efficient and autonomous mechanism is needed to map IoT services on the fog environment. The problem becomes more complicated when the various requirements of services including response time and cost are considered. With this motivation, we introduce a mixed-integer linear programming model to address the Fog Service Placement (FSP) problem in fog-cloud computing environments. This model is formulated as a multi-objective model with the aim of minimizing deadline violation, resource loss and service cost, and maximizing resource usage. Meanwhile, we propose an Elitism-based Genetic Algorithm to solve the FSP problem developed with a Shared Parallel Architecture (EGASPE). Our algorithm considers the efficient distribution of resources under certain Quality of Service (QoS) requirements and seeks a balance between exploration and exploitation. Also, EGASPE can save more resources for future requests by considering the deadline of IoT services over time. Extensive experiments have been performed to evaluate the performance of EGASPE on a synthetic fog environment. The obtained results show the effectiveness of EGASPE up to 6.5% on average compared to state-of-the-art algorithms.

A Proactive On-Demand Content Placement Strategy in Edge Intelligent Gateways

Bandwidth-intensive applications transmit large-scale video data in the network. It causes backhaul bottlenecks and affects user experience. Deploying edge cache on an access point (AP) is a popular method to bring content files closer to end-users, but it faces significant challenges, especially in efficiently predicting and satisfying different users' future content requests with limited cache capacity. In this paper, we propose an intelligent gateway assisted edge cache deployment strategy (GACD), which jointly considers traffic usage patterns in multiple APs and the impact of new content on the cache performance. In GACD, The cache content placement problem is formulated as a many-to-one bidirectional matching problem with a dynamic quota allocation, aiming to improve cache resource utilization and minimize the average delivery latency. To address this problem, we design a heterogeneous information networks based prediction algorithm to predict end-users' potential preference of new content files. Then, we adapt the seasonal autoregressive integrated moving average model for traffic usage prediction, and propose a many-to-one matching algorithm to achieve dynamic matching quota adjustment and efficient cache content placement. We conduct extensive real-world trace-based experiments to validate the performance of GACD. Compared with six alternative cache strategies, GACD improves the hit rate by 23.9% on average, reduces the average content delivery delay by 19.02%, and increases the accuracy by 31.02% on average.

Future Aware Pricing and Matching for Sustainable On-Demand Ride Pooling

The popularity of on-demand ride pooling is owing to the benefits offered to customers (lower prices), taxi drivers (higher revenue), environment (lower carbon footprint due to fewer vehicles) and aggregation companies like Uber (higher revenue). To achieve these benefits, two key interlinked challenges have to be solved effectively: (a) pricing -- setting prices to customer requests for taxis; and (b) matching -- assignment of customers (that accepted the prices) to taxis/cars. Traditionally, both these challenges have been studied individually and using myopic approaches (considering only current requests), without considering the impact of current matching on addressing future requests. In this paper, we develop a novel framework that handles the pricing and matching problems together, while also considering the future impact of the pricing and matching decisions. In our experimental results on a real-world taxi dataset, we demonstrate that our framework can significantly improve revenue (up to 17% and on average 6.4%) in a sustainable manner by reducing the number of vehicles (up to 14% and on average 10.6%) required to obtain a given fixed revenue and the overall distance travelled by vehicles (up to 11.1% and on average 3.7%). That is to say, we are able to provide an ideal win-win scenario for all stakeholders (customers, drivers, aggregator, environment) involved by obtaining higher revenue for customers, drivers, aggregator (ride pooling company) while being good for the environment (due to fewer number of vehicles on the road and lesser fuel consumed).

Anticipatory rebalancing of RoboTaxi systems

This study considers a RoboTaxi service, a futuristic transportation system whereby passengers can place transportation requests in real-time for a private ride. The passengers are picked up from their origins as soon as possible after placing a request and travel directly to their destination, from which the RoboTaxi may immediately continue to serve other customers, remain idle, or move to another location in the city in anticipation of future demand.The operation of RoboTaxi systems requires online decisions regarding vehicle dispatching and rebalancing. Dispatching assigns cars to requests. Rebalancing is the initiation of empty trips of idle cars to reduce future requests’ waiting time and take advantage of off-peak traffic hours. This study proposes an advanced method for rebalancing while assuming a simple yet effective dispatching policy.The optimal operation of a RoboTaxi fleet aims to minimize two competing values: the passengers waiting time between requests and pickups and the total distance traveled by empty vehicles. In a RoboTaxi system, vehicles may travel empty either when deadheading to pick up a passenger or while making rebalancing trips. There is a trade-off between the rebalancing effort, which includes various types of costs, and the passengers’ waiting time, representing the quality of the service provided by the system.

MAESTRO-X: Distributed Orchestration of Rotary-Wing UAV-Relay Swarms

This work details a scalable framework to orchestrate a swarm of rotary-wing UAVs serving as cellular relays to facilitate beyond line-of-sight connectivity and traffic offloading for ground users. First, a Multiscale Adaptive Energyconscious Scheduling and TRajectory Optimization (MAESTRO) framework is developed for a single UAV. Aiming to minimize the time-averaged latency to serve user requests, subject to an average UAV power constraint, it is shown that the optimization problem can be cast as a semi-Markov decision process, and exhibits a multiscale structure: outer actions on radial wait velocities and terminal service positions minimize the longterm delay-power trade-off, optimized via value iteration; given these outer actions, inner actions on angular wait velocities and service trajectories minimize a short-term delay-energy cost; finally, rate adaptation is embedded along the trajectory to leverage air-to-ground channel propagation conditions. A novel hierarchical competitive swarm optimization scheme is developed in the inner optimization, to devise high-resolution trajectories via iterative pair-wise updates. Next, MAESTRO is eXtended to UAV swarms (MAESTRO-X) via scalable policy replication, enabled by a decentralized command-and-control network augmented with: (1) spread maximization to proactively position UAVs to serve future requests; (2) consensus-driven conflict resolution to orchestrate scheduling decisions based on delay-energy costs including queuing dynamics; (3) adaptive frequency reuse to improve spectrum utilization across the network; and (4) a piggybacking mechanism allowing UAVs to serve multiple ground users simultaneously. Numerical evaluations show that, for user requests of 10 Mbits, generated according to a Poisson arrival process with rate 0.2 req/min/UAV, single-agent MAESTRO offers 3.8× faster service than a high-altitude platform and 29% faster than a static UAV deployment; moreover, for a swarm of 3 UAV-relays, MAESTRO-X delivers data payloads 4.7× faster than a successive convex approximation scheme; and remarkably, a single UAV optimized via MAESTRO outclasses 3 UAVs optimized via a deep-Q network by 38%.

What is the most cost‐beneficial way to prepare micafungin for paediatric oncology patients?

AbstractBackgroundA recent change in local paediatric antifungal guidelines resulted in an increase in micafungin use. Micafungin is an expensive medicine costing AUD $80 per vial. The pharmacy aseptic production unit (APU) has the ability to compound doses to share vials of micafungin to minimise wastage.AimTo determine whether aseptically compounded and individually prepared doses of micafungin would be more cost‐beneficial than nurses using whole vials prepared on the ward, and to determine which dose and numbers of doses would be most cost‐beneficial to compound.MethodA time and motion study of 20 patients determined how long it took for nurses and the pharmacy APU to prepare doses. An economic analysis of time, cost of labour, and consumables outlined the cost to make one to seven doses with doses ranging from 10 to 200 mg.ResultsOn average, it took the pharmacy APU 42 min and nurses 9 min to prepare a single dose. Aseptically compounding one dose, whole vials (doses of 50 mg increments), and doses near whole vials had no cost‐benefit. Aseptically compounding five doses or more had would result in a cost‐benefit.ConclusionDepending on the dose required, cost savings could be obtained when the pharmacy APU prepared multiple doses, with a maximum saving of $350.10 for a 7‐day course. Determining baseline costs of micafungin preparation is relevant if there are future requests to compound medicines as the data could be extrapolated and applied to the new request.

DMACN: A Dynamic Multi-Attribute Caching Mechanism for NDN-Based Remote Health Monitoring System

Named Data Network (NDN) advocates the philosophy of accessing IoT data owing to its location independence feature. This enables routers to pre-cache content and serves the future requests for the same content on a local basis. Such architecture demonstrates huge application potential in the E-health field. In order to achieve efficient healthcare treatment and administration for both patients and medical professions, the optimization of storing patients&#x0027; real-time, large-scale physical data is of necessity. We propose a <u><b>D</b></u>ynamic <u><b>M</b></u>ulti-<u><b>A</b></u>ttribute <u><b>C</b></u>aching mechanism for <u><b>N</b></u>DN-Based remote health monitoring system (DMACN). In our model, we adopted a predictable consumer-driven freshness mechanism with low computation cost to satisfy the freshness-sensitive nature of health data. A novel content popularity model based on Analytic Hierarchy Process (AHP) and medical-grade parameters are proposed to handle the doctor-decision-making-coupled Interest sending mechanism of a remote health monitoring system. The final simulation results show DMACN has strong robustness against intensive requesting and complex contents. It is also shown that its performance surpasses existing mechanisms. The <i>Cache Hit Ratio</i> exceeds 37.5&#x0025; to FIFO and LRU, 220&#x0025; compared with CPFC, and 55&#x0025; for CTDICR for both consumer and producer tests. On the other hand, the <i>Latency</i> is about 23.8&#x0025; lower than FIFO and LRU, 46.6&#x0025; lower compared with CPFC, and 35.4&#x0025; for CTDICR.

A smart system combining real and predicted data to recommend an optimal electric vehicle charging station

The electric vehicle (EV) is considered an attractive alternative to a conventional vehicle, due to its potential beneficiation in decreasing carbon emission. But the battery range anxiety is a key challenge to its wide adoption and also the EV drivers spend so much time in public charging stations (CS) to charge especially during peak times. In this paper, we propose a charging station selected system (C3S) to control and manage EVs charging plans. Moreover, the C3S system proposed consists of a set of algorithms that are proposed to recommend a suitable CS for EV charging requests. The CS selection is based on minimizing travel time and takes into account in real-time the queuing time at CS, EVs' charging reservations, and the predicted time of EVs' future charging requests. Besides, we proposed three different strategies for predicting the EVs incoming and controlling the uncertainty matter of the dynamic arrival of EV charging requests. As part of the Helsinki City scenario, the evaluation results demonstrate the performance, especially at peak times, of our proposed C3S with regard to the CS recommendation which has the minimum total trip duration.

Data-driven customer acceptance for attended home delivery

Home delivery services require the attendance of the customer during delivery. Hence, retailers and customers mutually agree on a delivery time window in the booking process. However, when a customer requests a time window, it is not clear how much accepting the ongoing request significantly reduces the availability of time windows for future customers. In this paper, we explore using historical order data to manage scarce delivery capacities efficiently. We propose a sampling-based customer acceptance approach that is fed with different combinations of these data to assess the impact of the current request on route efficiency and the ability to accept future requests. We propose a data-science process to investigate the best use of historical order data in terms of recency and amount of sampling data. We identify features that help to improve the acceptance decision as well as the retailer’s revenue. We demonstrate our approach with large amounts of real historical order data from two cities served by an online grocery in Germany.

Energy-optimal DNN model placement in UAV-enabled edge computing networks

Unmanned aerial vehicle (UAV)-enabled edge computing is emerging as a potential enabler for Artificial Intelligence of Things (AIoT) in the forthcoming sixth-generation (6G) communication networks. With the use of flexible UAVs, massive sensing data is gathered and processed promptly without considering geographical locations. Deep neural networks (DNNs) are becoming a driving force to extract valuable information from sensing data. However, the lightweight servers installed on UAVs are not able to meet the extremely high requirements of inference tasks due to the limited battery capacities of UAVs. In this work, we investigate a DNN model placement problem for AIoT applications, where the trained DNN models are selected and placed on UAVs to execute inference tasks locally. It is impractical to obtain future DNN model request profiles and system operation states in UAV-enabled edge computing. The Lyapunov optimization technique is leveraged for the proposed DNN model placement problem. Based on the observed system overview, an advanced online placement (AOP) algorithm is developed to solve the transformed problem in each time slot, which can reduce DNN model transmission delay and disk I/O energy cost simultaneously while keeping the input data queues stable. Finally, extensive simulations are provided to depict the effectiveness of the AOP algorithm. The numerical results demonstrate that the AOP algorithm can reduce 18.14% of the model placement cost and 29.89% of the input data queue backlog on average by comparing it with benchmark algorithms.

DRL-Assisted Reoptimization of Network Slice Embedding on EON-Enabled Transport Networks

5G transport networks will support dynamic services with diverse requirements through network slicing. Elastic Optical Networks (EONs) facilitate transport network slicing by flexible spectrum allocation and tuning of transmission configurations. A major challenge in supporting dynamic services is the lack of priori knowledge of future slice requests. As a consequence, slice embedding can become sub-optimal over time, leading to spectrum fragmentation and skewed utilization. This in turn can block future slice requests, impacting operator revenue. To address this issue, operators can periodically re-optimize slice embedding for reducing fragmentation. In this paper, we address this problem of re-optimizing network slice embedding on EONs for minimizing fragmentation. The problem is solved in its splittable version, which significantly increases problem complexity, but also offers more opportunities for a larger set of re-configuration actions. We employ simulated annealing for systematically exploring the large solution space. We also propose a greedy algorithm to address the practical constraint of limiting the number of re-configuration steps. Moreover, we present a novel method based on Deep Reinforcement Learning (DRL) for determining when performing re-configuration is most effective. Our extensive simulations demonstrate that the greedy algorithm yields a solution very close to that obtained using simulated annealing while requiring orders of magnitude lesser re-configuration actions. Finally, we show that by applying the greedy algorithm periodically on the network according to the DRL-based time selection algorithm, a significant improvement in the total number of accepted slice requests can be achieved with only performing a limited number of re-configuration operations.

DSPPV: Dynamic service function chains placement with parallelized virtual network functions in mobile edge computing

This study configures an architecture based on Deep Reinforcement Learning (DRL) with the aim of providing online services to end users in Mobile Edge Computing (MEC) networks. Network Function Virtualization (NFV) technology can provide these services in MEC by turning hardware middleboxes to Virtual Network Functions (VNFs) for mobile users. Network services are chained as an ordered sequence of VNFs named Service Function Chains (SFCs), which are provided by directing traffic to the required VNFs. Meanwhile, the SFC placement problem is challenging for provisioning online service requests under limited resource availability and improving quality of service. We propose DRL-based Dynamic SFC Placement method with Parallelized VNFs (DSPPV) to solve this problem that seeks to maximize long-term expected cumulative reward. By sharing VNFs in parallel, DSPPV can achieve computational acceleration in providing online services. Any pair of VNFs that do not conflict on traffic can process the packet simultaneously, so configuring SFC with parallelized VNFs will reduce the SFC length and thus reduce the latency. In addition, DSPPV increases the ability to process future requests by extracting the distribution of initialized VNFs. The performed simulations show the effectiveness of the proposed architecture. Specifically, the obtained numerical results show that the average number of accepted requests has improved between 4% and 15%.

A model for green order quantity allocation in a collaborative supply chain

Several models have been proposed in the literature for allocating quantity orders, but few of them consider collaboration to enhance supply chain (SC) performance sustainably. Despite the increasing attention given to sustainability and the triple bottom line (TBL) criteria, the existing models for order allocation are limited to considering TBL scores only in the phase of selecting suppliers, thus evaluating mainly environmental attributes, without integrating economic and social ones. For these reasons, based on the main research gaps in the current literature, this paper aims to investigate the green order allocation problem in an SC with high operational variability, exploiting the advantages of a collaborative and sustainable SC. The model we propose considers the uncertainty of the demand, the limited capacity of the suppliers, the minimum levels of purchase and production, the environmental, economic, and social scores of suppliers, and the total emissions related to the production, ordering, maintenance, and transport of the allocated items. Our model is based on allocation curves that are continuously updated and shared between the customer and the supplier. On the customer side, it allows to dynamically optimise order allocation, and on the supplier side, it provides clear information on the possible future requests from the customer based on the overall and real economic, environmental, and social performance. To validate the model, the authors applied it to a numerical example, which resulted in minimising both costs and the TBL impacts. Moreover, considering the TBL scores only in the selection of suppliers led to non-optimal choices in the present case study; for example, assigning the largest share of the order to the supplier with the lowest TBL score did not result in the most sustainable allocation. The continuous exchange of information and the synchronism triggered by the buyer–supplier collaboration allows managers to know in advance of any critical issues that may arise in the future and intervene promptly to reorganise the allocation. Therefore, the proposed model can help both researchers and practitioners establish a systematic approach to green order allocation and might help firms increase their sustainability degrees throughout the supply chain.

A fuzzy delay-bandwidth guaranteed routing algorithm for video conferencing services over SDN networks.

Video conferencing is one of the advanced technologies for users that allows online communication despite long distances. High quality communication and ongoing support for the principles of video conferencing service that can be achieved through Software-Defined Networking (SDN). SDN is a new architecture for computer networks that separates the control plane from the data plane to improve network resources and reduce operating costs. All routing decisions and control mechanisms are made by a device called a controller. Traffic engineering can be well implemented in SDN because the entire network topology is known to the controller. Considering SDN features, user requests can be dynamically routed according to current network status and Quality of Service (QoS) requirements. In general, the purpose of SDN routing algorithms is to maximize the acceptance rate of user requests by considering QoS requirements. In this literature, most routing studies to provide satisfactory video conferencing services have focused solely on bandwidth. Nevertheless, some studies have considered both delay and bandwidth constraints. In this paper, a Fuzzy Delay-Bandwidth Guaranteed Routing (FDBGR) algorithm is proposed that considers both delay and bandwidth constraints in routing. The proposed fuzzy system is based on rules that can postpone requests with high resource demands. Also, the purpose of the FDBGR is to distribute the network workload evenly for all requests, where this is done by maintaining the capacity to accept future requests. The combination of conventional routing algorithms and SDN provides remarkable improvements in mobility, scalability and the overall performance of the networks. Simulations are performed on different scenarios to evaluate the performance of the FDBGR compared to state-of-the-art methods. Besides, FDBGR has been compared with a number of most related previous works such as H-MCOP, MH-MCOP, QoMRA, QROUTE and REDO based on criteria such as number of accepted requests, average path length, energy consumption, load balancing, and average delay. The simulation results clearly prove the superiority of the proposed algorithm with an average delay of 48ms in different topologies for video conferencing applications.

A Fuzzy Based Routing Approach for Improving Online Conferencing Services in Software Defined Networking

In recent years, the provision of applications based on conferencing services over Software-Defined Networking (SDN) is increasing. So far, extensive studies have been conducted on improving conferencing services and supporting high-quality communication in SDN, and the importance of the issue has been understood by the research society. One of the challenges of guaranteeing conferencing services in SDN is the routing of network traffic taking into account Quality of Service (QoS) requirements. Most routing techniques in SDN guarantee end-to-end delay and bandwidth constraints. However, resource loss also plays a role in network performance, where it is clearly neglected in existing works. This study presents a guaranteed Latency-Bandwidth-Loss Routing algorithm based on Fuzzy approach to improve the performance of conference services in SDN (LBLRF). The purpose of the fuzzy-based approach is to increase the number of routed requests and improve the performance of conference services in SDN. LBLRF uses a postponement mechanism to improve the conference service, where it gives higher priority to the requests with low resource demand for connection. In addition, LBLRF is equipped with a hold time mechanism to release the reserved resources after satisfying the request requirements. This mechanism can increase the processing capacity of future requests by conserving network resources. Evaluations have been done by defining several scenarios on a synthetic SDN environment. The results of the experiments show that LBLRF provides a better request acceptance rate between 7% and 16% compared to the best results of existing algorithms.