Search

The main goal of Random Early Detection (RED) is to avoid traffic congestion. Since network requirements differ significantly, it is not an optimal approach to establish RED parameters with a constant value. We are proposing a new model Congestion Control Algorithm using Buffer Occupancy RED (CCA-BO-RED) with the dynamic tuning of RED parameters. We measure the rate of occupancy of the queue for this reason and treat it as a congestion parameter that will be predicted when the queue is crowded. The proposed new metric, that is called Buffer Occupancy (BOC), infers network conditions from the rapidity of the buffer occupancy in the router. A new range variable has been introduced to improve the performance of the priority queue of the existing RED-based algorithm which improves the overall performance of networks. For each packet, the minimum and maximum threshold has been updated and dropped with probability (Pa) for a special condition. Instead of multiplicative increase and decrease the maximum probability, the scheme uses additive-increase and multiplicative-decrease. The simulated results prove that our approaches are better than RED in terms of throughput, end-to-end delay, packet delivery ratio and goodput.

The present work is intended to provide a solution through the development of a chatbot using the Watson assistant and Node-RED UI for booking movie tickets. It ensures error-free booking of tickets without human agent intervention smoothly even during peak hours. It saves time and makes the process much more easy for the customer as well as for the service provider by eliminating the elongated process starting from waiting in the queue of booking counter to getting the answers to repetitive queries. A web-based conversational flow is generated by only clicking on an URL from any electronic gadget with internet access.

The basic philosophy behind RED is to prevent congestion. To overcome the problem in RED without changing queue weight parameter, we have proposed two approaches to control the congestion by introducing a range parameters and control mechanism which will belong between minimum and maximum threshold. The current queue size is controlled together with average queue size. A new range variable has been introduced to improve the performance of priority queue of existing RED algorithm which improves the overall performance of networks. For each packet, minimum and maximum threshold has been updated as MaximumThreshold = 2 * MaximumThreshold and MinimumThreshold = (2 * MaximumThreshold + 3 * MinimumThreshold)/5 + MinimumThreshold. For the condition (MinimumThreshold MinimumThreshold) then each arriving packet is dropped with probability Pa. As the AVG queue length approaches the minimum threshold value, our method automatically adjusts queue parameters to match the queue conditions, reducing queuing time and increasing throughput. The simulated results proof that our approaches are better than RED in terms of throughput, end to end delay, packet delivery ratio and goodput..

This paper examines the optimality of a single server queues where the server is permitted to take two differentiated hiatus. The inter-arrival times of arriving clients, the service times and two hiatus times are all exponentially distributed with \(\lambda \), \(\mu \), \(\alpha _1\) and \(\alpha _2\), respectively. The ceiling of clients admitted into the system is of L. The stationary system size distributions of the model by using probability generating functions are obtained. Optimization of the model is studied using particle swarm optimization. A few numerical arguments validating the impact of parameters pertaining to in our system on vital performance measures of the model are hosted.

The real-time systems have a set of processes continuously running in the system and will also have a set of processes triggered on events or user interaction. In such systems, the number of processes in scheduling queue varies based on frequency of trigger activation. The hard real-time systems are designed to meet the deadline of all processes under peak load of processes. Hence, the execution speed of the processing unit is chosen much higher than the required speed for average load. The DVFS technique addresses the issue based on process load; however, it does not take deadlines into account while scaling the processors operating frequency.

In a supermarket or a mall, people come to purchase products and during the time of payment, they need to calculate and know about the total bill which is hectic in nature. In order to overcome this problem, an application is created which keeps track of transaction history of both past and current billing records. This project is done to simplify shopping methods and reduce the long queue during the process of billing. In the previous models, authors have failed to make use of applications for shopping, also the previously proposed models had RFID scanner in every trolley for reducing the queues which was more expensive. So basically, there was no application created for shopping and and an alternate way for scanning the products other than the RFID scanner was not introduced in malls. The methodology used here consists of a centralized system for the recommendation and online transaction. Devices that are used in this prototype are a laptop with webcam and load cell. The final product can be replicated using Raspberry Pi. The objective of this model is to have an application that can scan the products and even register the new user.

Single Non-perishable product stochastic inventory system with maximum capacity of S items is served by a single server, which serves Israeli queue. Reorder is placed only after the level of inventory reaches zero. Customers form an Israeli Queue to get service with at most N different groups with the consideration of unrestricted batch size service. Service time is independent of batch size. Markov structure of the model is identified, and state probabilities are arrived by Numerical approach using MATLAB coding. Also arrived the performance measures like first-order product density of replenishment, mean number of groups in the System, System throughput depends on Inventory level, mean number of groups that are bypassed by the arriving customer. Model validated by numerical illustration.

In the health service center, many people often meet in a queue and tend to be chaotic, causing uncomfortable conditions and time inefficiency. From these problems, it was thought that technology was needed to make it easier for users to register and monitor queues. In previous studies, we have successfully implemented a digital queue system using a mini PC as the main control. This system has been implemented at the public health center “Puskesmas” Bojong Soang, Bandung. However, the mechanism of queuing, registration, or ticketing and monitoring could not be done online in an application. Therefore, in this study, an Android-based application was developed to register and monitor queues online and in real-time. The main purpose of this study is to build an integrated system between existing queuing machines and mobile applications so that the queuing mechanism can be done online. The proposed system consisted of three main components, namely the digital queue machine, server-cloud, and client application in the user’s smartphone. All modules were connected via the cloud with the developed protocol. This application was called “Q-Puskesmas”, which could run on the Android mobile platform with a minimum version of 5.0. The proposed system is able to display location identity, the latest queue number, and a menu for registering. From the test results, the application can update the queue number in real-time according to the existing conditions with a delay of

The Internet of things is evolving in everyday life by partner conventional articles into gathering. For example, in a general store, all things can be related to each other, modeling a splendid shopping system. In such an Internet of things system, an efficient radio frequency identification tag can be associated with each thing which, when put into a shopping cart, can be thus examined by a truck outfitted with a radio frequency identification for every client. In like manner, charging can be coordinated from the shopping bushel itself, shielding customers from holding up in a long queue at the checkout. Moreover, sharp racking can be incorporated into this system, outfitted with radio frequency identification per clients, and can screen stock, perhaps in like manner invigorating a central server. Another favorable position of this kind of structure is that stock organization ends up being significantly less complex, as all things can be normally scrutinized by a radio frequency identification for every client as opposed to physically analyzed by a specialist.

The audio transmission over WLAN requires the better Quality of Service (QoS). In this paper, we are proposing and implementing the novelty adaptive mapping of cross-layers scheme for height the Quality of Services for transmitting the scalable video streams in IEEE 802.11ac wired less environment tends to approximating of the AC’s queue length and delaying of the timely constraining of the streams for which it is to be transmitted. The results clearly states that performing of propose dynamic cross-layers mapping is better with respect to the mapping of static cross-layers scheme.

It is generally observed that billing at shopping stores takes a lot of time, especially during holidays and weekends. A person in India visits a grocery store 1.6 times every 15 days and spends nearly 43 min there. Therefore, it is necessary to have an efficient mechanism, which would help people to shop in smarter way. With this system, we have come up with a solution to the problem of long queues while shopping in a mall or a store. So, our solution uses the concept of Radio Frequency Identification. It is necessary to have RFID tags (transponder) embedded on all the products. So, when the purchased product comes in the range of antenna, the passive tag’s circuit gets activated and sends the prestored data. On the other side, reader (transceiver) will receive these data and send it to the computer. This helps in establishing a very efficient way of billing significantly helping customers save a lot of time. The aim of this solution is to put an end to the long queues we encounter at the shopping centers. The customer has to just drop in his purchases at the RFID zone and head to the billing counters for payment thus establishing a very unique and efficient shopping practice, which is much needed in these modern times.

Adaptive bitrate (ABR) has been widely used in video streaming to adjust the bitrate of video chunks and optimize user’s quality of experience (QoE). In this paper, we propose a novel client-side ABR algorithm, called Dynamic Buffer Control (DBC), by taking both bitrate smoothness and rebuffering into consideration. We formulate the QoE optimization problem by taking the perceived bitrate as the utility while taking bitrate smoothness and rebuffering as penalty functions. To solve this problem, we propose virtual queues to translate penalty functions’ constraints into queueing problems and use Lyapunov Drift Optimization to achieve maximum bitrate and stable queues. We evaluate the performance of DBC and compare it with both model-based and learning-based ABR algorithms under multiple QoE metrics and trace datasets. Evaluation results show that DBC outperforms model-based algorithms, i.e., MPC, BOLA and FESTIVE with 13%, 20% and 27% gain respectively. DBC outperforms learning-based ABR schemes, i.e., Comyco and Pensieve, with 4% and 11% gain respectively.

This paper proposes a method for dynamically forming teams and assigning appropriate tasks to their members to provide services accomplished by groups of agents of different types. Task or resource allocation in multi-agent systems has drawn attention and has been applied in many areas, such as robot rescue, UAV wireless networks, and distributed computer systems. The proposed method allows agents to belong to more than one team simultaneously for efficiency based on the reciprocal trust relationship, which reflects the past performance of cooperative work, and thus allows each agent to have a queue to undertake multiple tasks. In such a setting, in addition to the communication time, the tasks in the queue can even cause processing delays, leading to instability in the observed information from the leader who selects the team members. Our experimental evaluation shows that the proposed method can efficiently enable stable team formation even in this situation. We also analyze the reasons for this efficiency.

This article aims to focus a study under different scenarios on retrial G-queues. In retrial queues, a customer makes an attempt, again and again, to avail the service on being rejected by the server whereas, in G-queues, a negative customer arrives only when the server is occupied with a positive customer and forces it to leave the system thereby causing an interruption in the service. From the past few decades, a lot of researchers are being attracted towards this field. Our investigation includes all the major research articles from reputed journals which have been published till now. The work is divided on the basis of different models and methodologies of queueing analysis along with various applications. Discrete-time retrial queues with negative customers are also discussed. Our main focus is to help and motivate all researchers who are willing to explore retrial queueing theory with negative customer arrival.

In this chapter, we propose a clustered Virtual Machine (VM) allocation strategy based on a sleep-mode with wake-up threshold to achieve green computing. The VMs in a cloud data center are clustered into two modules, namely, Module I and Module II. The VMs in Module I remain awake at all times, while the VMs in Module II go to sleep under a light workload. We build a queue with an N-policy and asynchronous vacations for partial servers to capture the stochastic behavior of tasks with the proposed strategy. We derive performance measures of the system in terms of the average latency of tasks and the energy saving rate of the system, respectively. Furthermore, we present numerical results to demonstrate the impact of the system parameters on the system performance. Finally, we construct a system cost function to trade off different performance measures and develop an intelligent searching algorithm to jointly optimize the number of the VMs in Module II, the wake-up threshold and the sleep parameter.

Transmission control protocol (TCP) is an important transport layer protocol. Some survey reports predict that 85% of the network traffic is carried out by TCP. UDP, which is non-responsive to the network traffic, has a share of 15%. But, because of increasing popularity of real-time and multimedia applications, the transportation share of UDP is reasonably increasing. In hybrid networks, the traffic type is always heterogeneous. Hence, there is a need to understand the effect of UDP flows on TCP flows. In this paper, a model-based analysis of TCP performance in presence of UDP flows is presented. CBR flow uses UDP application. Round trip delay, queue dynamics, probability of packet drop, congestion window dynamics and throughput performance are analyzed in the absence of UDP and in the presence of UDP flows. The experimental results reveal that performance of TCP decreases by 40% in presence of UDP flows. This is because UDP captures large amount of network resources and make TCP deprived of the resources. The experimental results say that there is a need to tune UDP such that it is responsive to the network traffic. Simulation was conducted using MATLAB tool. Analysis of experimental results has been made using statistical data and graphs.

Considering the property of self-similar traffic shown widely in the networks with multimedia transmission, in this chapter, we present a method to analyze the performance of the enhanced power saving class type III with self-similar traffic. According to the operating principle of the sleep mode in the enhanced power saving class type III, considering the self-similar nature of massive multimedia packets in wireless mobile networks, we build a discrete-time batch arrival multiple vacation queueing model with vacation-delay, in which the batch size is supposed to be Pareto distributed. We present the boundary state variable theory for the batch arrival vacation queueing model to show queue length, waiting time and busy cycle in the steady state. Moreover, we derive performance measures of the system in terms of the handover rate, the energy saving rate of the system, the system utilization and the average response time of data packets, respectively. Finally, we present numerical results to demonstrate the influence of the system parameters on the system performance with different offered loads and different degrees of self-similar traffic.

When using a Discontinuous Reception (DRX) mechanism in Long Term Evolution (LTE) for wireless communication of high-speed data, two different operational modes are employed: Idle DRX and Active DRX. In this chapter, we propose an enhanced energy saving strategy based on the Active DRX mechanism in an LTE system to improve the sleep strategy for a better balance between response performance and energy efficiency by introducing a sleep-delay timer. We build a discrete-time multiple-vacation queueing model with a vacation-delay period and a set-up period by addressing the busy period, the queue length, the waiting time and the busy cycle. We derive performance measures of the system in terms of the handover rate, the energy saving rate of the system and the average response time of data packets, respectively. We present numerical results to show the impact of the thresholds of the short DRX stages, the time lengths of the sleep-delay timer, the short DRX stage and the long DRX stage on the system performance so that we can evaluate the influence of the configuration parameters on the system performance. Finally, by considering the trade-off between different performance measures, we optimize the enhanced energy saving strategy for the Active DRX.

This paper presents a method for particle filter (PF)-based joint state and parameter estimation for a stochastic hybrid model of queue-length dynamics for urban traffic. The model represents continuous variables, including the stochastic traffic flow rates entering and exiting the queue during successive green and red periods, and discrete event variables including the green/red switching and the traffic flow modes (free flow versus congested). A particle filtering approach is used for jointly estimating the transition matrix of the traffic modes, the parameters of the autoregressive (AR) model of the traffic flow rates, as well as for estimating and predicting the traffic flow rates and the queue length themselves. Obtaining computationally efficient PFs for this requires several improvements to existing techniques. We use an optimal tuning kernel smoothing approach to estimate the parameters of the first-order AR model in combination with the Dirichlet distribution approach to update the parameters of the transition probability matrix of a first-order Markov chain. The proposed technique is validated and evaluated by comparing the queue size estimations with synthetic data generated by a VISSIM traffic micro-simulator. The results of this analysis confirm that the hybrid model together with the PF parameter estimator gives satisfactory results, properly capturing the evolution in time of queue length and traffic flow. This algorithm may therefore be included in the feedback loop of adaptive traffic signal controllers.

A restaurant is a public place, and it is open to everyone to take a drinks and food against money. The restaurant offers service of food and drinks desires to satisfy the customer. At the moment, the payment process at restaurants is still manual and ineffective because it uses a cash record. A cashier will check what foods are required, then calculate them with the cash record, and this process will take a long time and be ineffective. In this paper, design application for Iraqi food detection and also reduces long queues at certain hours in restaurants and estimate food costs automatically applying the convolutional neural network (CNN) classification method. The results of the application are can detect six types of food on 80% data partition training data and 20% test data, with the image taken in storage. The accuracy of the appley of CNN reaches 100% with a detection time of 3.59 s.

The ground movement is one of the most critical airside operations. It includes two sub-problems: routing and scheduling and serves the purpose of guiding aircraft on the surface of an airport to meet the departure schedule while minimizing overall travel time. To achieve that purpose, ground movement controllers manage the taxi-route assignments and taxi-time estimation for each aircraft in arrival or departure queue. A high-accuracy taxi-time calculation is required to increase the efficiency of airport operations. In this study, we propose a data-driven approach to construct features set and build predictive models for taxi-time prediction for departure flights. The proposed approach can suggest the taxi-route and predict the corresponding taxi-time by analyzing ground movement data. The controller’s operational preferences are extracted and learned by machine learning algorithms for predicting taxi-route and taxi-time of given aircraft. In this approach, we take advantage of taxiing trajectories to learn the controller’s decision, which reflects how the controller had decided the routing for a given situation. Two machine learning models, random forest regression, and linear regression are implemented and show similar performances in estimating the taxi-time. However, since the random forest is an ensemble method that has advantages in handling outliers, performing feature selection, and assessing feature importance, it can provide more stable results and interpretability, for real operations. The predictive model for taxi-time can predict the taxi-out time with high accuracy with given assigned taxi-route. The model can cover the controller’s decision up to 70% in the top-1 and 89% in top-2 recommends. The mean absolute error is less than 2.07 min for all departure flights, and root mean square error is approximately 2.5 min. Moreover, the ± 3-minute error window can cover around 76% of departures, while more than 95% of departures are within the ± 5-minute error window.

This chapter aims to develop a mathematical queueing model for multi-phase essential repair (MPER) systems with unreliable servers. Moreover, the service of the customer is interjected due to unavailability of the server. The failed items get repaired following MPER. Multiple working vacations (MWV) have been chosen by the servers in case of no availability of customers in the system. As the first vacation completes, the servers join the system and provide service to awaited customers. If no more awaited customers are presented in the system, server may continue their other vacations. The joining of customers to MPER system follows the Poisson distribution. Time-dependent probabilities are calculated by Runge–Kutta (RK). With the help of MATLAB software, numerical results are derived for the MPER system. Comparative results are also provided using adaptive network-based fuzzy inference systems (ANFIS). At last, the conclusion is given.

This paper proposes a priority-based optimized scheduling (PBOS) algorithm for cloud-based big data applications. In this algorithm, for each incoming task request, the task size and expected completion time are estimated. At the same time, for each VM on a host, the processing capability and current load are estimated. A priority queue for each task is then created in which the tasks are stored in the ascending order of task size and completion time. Similarly, a priority queue for VMs is also made in which the VMs are stored in the ascending order of current load and processing capacity. Then, the scheduling algorithm intelligently maps each user tasks from task queue to the suitable matching VM from VM queue. The process is continued until both these queue become empty. By experimental results, it is shown that PBOS achieves lesser response time and execution time.

In this paper, multi-objective optimization is used to solve the signal synchronization problem in arterial traffic roads, where a traffic dispersion module is introduced to further expand the solution space. By incorporating the models of delay time, queue length and stop times into the optimization, a first model called M1 is established. In the second model M2, the free flow speed assumption is replaced by a traffic dispersion module for better estimating the link travel time. A simulation study is then carried out on an arterial road, and the results show that the proposed strategy improves the performance of the traffic system compared to the current timing scheme and M2 has the best performance among all solutions in this paper, and the delay is reduced for about 24%.

In this paper, we study the problem of assignment of suitable berths to the vessels under different scenarios of vessel berthing policies, priorities and vessel serving or container handling-off. The problem was solved as a queuing system with non-preemptive priority. The objective was to maximize the utilization of the berth under different service levels. Different scenarios for berthing process of vessels and unloading of containers at the container terminal were considered to evaluate the performance of the system and to obtain the optimal service level parameters. The model considered a system in which different types of containers were managed together in an integrated manner. The steady-state behaviours of the expected waiting time, queue length, server utilization rate and the optimal number of servers necessary to attain given service levels for the different container types were studied experimentally under different conditions of arrival and service rates.

In order to reduce the average response time of data packets while guaranteeing a greater energy saving rate of the system, in this chapter, we propose an enhanced Active Discontinuous Reception (DRX) mechanism with a sleep-delay strategy in the Long Term Evolution (LTE) system to influence the downlink transmission at the User Equipment (UE). Utilizing several logical channels for one connection, we build a multiple synchronous vacation queueing system with a wake-up period and a sleep-delay. We derive performance measures of the system in terms of the energy saving rate of the system, the blocking rate and the average response time of data packets. We present numerical results to show the validity of the proposed enhanced Active DRX mechanism with a sleep-delay strategy. Finally, by constructing a system profit function, we optimize the number of the logical channels for one connection, the time lengths of the sleep-delay timer and the sleep period.

This paper demonstrates the ticket allocation problem in an IT operations incident management environment, by infusing a data-driven approach using data analytics and rule-based machine learning techniques. While conventional methods such as cherry-picking of tickets, round-robin methods exhibit bias in terms of the perceived ability of human agents, an analytically suave data-driven approach eliminates such bias, while delivering substantial performance benefits. The ticket allocation problem deals not only with how many tickets to allocate to each resource but also what job to allocate to which resource. Thus, solving two potential riddles baffling the Operations Managers, i.e., allocation as well as assignment. In turn, our solution optimizes the queue waiting time which can reduce the cost and increase customer satisfaction. The solution to the ticket-queuing problem can particularly help in ticket prioritization. Thus, we have dissected the ticket allocation problem in three parts and have attempted to provide solutions to all the three aspects. We assimilated and fused the three-pronged solution into a single solution that can be readily used not only in IT operations but in any operations environment where ticket allocation needs to be optimized. In this paper, we are solving the real-time ticket allocation problem, through agent-scoring approach. This depends on agent success rate and agent effort hours. This is an instantaneous solution that will work in real time, as soon as a ticket lands on the floor.

In this paper, a priority-based modified deficit weighted round-robin (PMDWRR) traffic forecast system is prposed that could assure the actual precedence of consumed traffic. The projected forecast system is a distinction of the changed modified deficit weighted round-robin (MDWRR) system for Gigabit Passive Optical Network (GPON) and pledges the maximum precedence consumed traffic communication with delay minimization. WRR procedure to be conscious of bandwidth and expands the impartiality. But for convinced traffic categories, impartiality is not the anticipated behavior. To realize expected provision for subtle, actual traffic, a precedence level for arrangement desires to be announced. By empowering stern precedence, or by contributing to numerous precedence stages and by means of DWRR arrangement among queues with the similar precedence stages, provision pledge with respect to interruption and damage guard can be attained for demanding traffic categories, such as voice and actual distribution.

The most important aspect of Broadband Wireless Access (BWA) networks is to provide heterogeneous traffic flows with a guarantee of Quality of Service (QoS). To this end, we have designed an efficient adaptive Weighted Hybrid Queue Scheduling Scheme (WHQSS) to support differential QoS requirements for traffic classes, which are heterogeneous in BWA networks. At first, the performance of WHQSS is analyzed in comparison with two important Queue Scheduling Schemes (QSSs), namely Priority Queue Scheduling Scheme (PQSS) and Hybrid Queue Scheduling Scheme (HQSS) those are designed with some major existing scheduling algorithms. A single analytical platform has been created using 3D Continuous Time Markov Chain (CTMC) model to investigate the performance in terms of several parameters like mean number of packets waiting in the queue, throughput, mean queuing delay, packet loss probability, and fairness index of the said QSSs. The comparative performance results clearly reveal the superiority of WHQSS providing optimal QoS guarantee and highest fairness than other QSSs for heterogeneous traffic classes.

The scheduling algorithms for IEEE 802.16 standard are designed with the predominant goals of throughput optimization, ensuring fairness and Quality of Service (QoS) provisioning. In this work, enhancements are proposed to the existing Weighted Fair Queuing (WFQ) and Deficit Weighted Round Robin (DWRR) scheduling algorithms to efficiently utilize the unused units. In WFQ, additional units may be assigned to a queue, thus reducing the service rate. Instead in Enhanced WFQ (EWFQ), multiple queues are served in a round by effectively utilizing the unexploited units. In DWRR, a queue is not serviced if the size of the packet at the front of the queue exceeds the available quantum. Enhanced DWRR (EDWRR) checks for packets with sizes less than the Deficit Counter (DC), sorts the queue and services a smaller packet in the current round. Further, if the queue that is currently served becomes empty, the DC is transferred to the ensuing active queue instead of making it zero. This helps in servicing more number of packets in a round. The proposed scheduling schemes are proficient in servicing specific traffic flows.

Consider various discrete-time systems where the time axis is divided to put something into a particular space that is designed for it, known as spots. In view of the services and evacuation could be taken as spot boundaries, where their periods are integral multiple of spot span. In general, distributed service time is denoted by Geo/G/1 and single-server system with discrete time has Bernoulli arrival process. A Geox/G/1 system is the basis of all the extended models. Such particular points are in fact identical with the distribution of the system states observed at an arbitrary point on the continuous-time domain. This leads a random point which lies in the middle of a spot on the continuous-time domain with probability 1.

Cloud technology is a fast-growing technology which supports wide range of services with less cost. The majority of the costly services like ERP software tools, hardware equipments, and packages can be availed here with the basic concept of ‘What and How much we used, per for that only.’ There are different concerns to be considered in cloud computing for researchers; they are load balancing, security issues, and availability of resources. Join idle queue (JIQ) is a dynamic algorithm used for virtual machine allocation for incoming jobs within a datacenter. In that JIQ algorithm, some issues were found. In this paper, we proposed enhancements in JIQ algorithm and used in optimized load balancing across the virtual machines in the data center.

Real-time pricing in smart grid is an ideal power pricing mechanism, which has advantages such as peak shaving, energy saving, environmental protection, and the minimum cost of the power supply company. The rigid demand is not affected by the price, and the elastic demand is affected by the price. The elastic demand is random. Therefore, this paper is devoted to stochastic elasticity of electric demand in smart grid with real-time pricing algorithm. We adjust the user’s elastic demand through the real-time price, which also makes the grid load balancing, so as to achieve the grid utility maximization (GUM). We establish the stochastic elasticity of electric demand in smart grid model, design a subgradient dual decomposition algorithm based on Lyapunov to solve the problem. The simulation result shows that the algorithm is able to achieve the stable and load balancing of the power demand and queue and achieve the maximum utility of power grid.

Assurance of quality of service (QoS) to secondary users in cognitive radio networks is a challenge as the secondary users have to use the spectrum on an opportunistic basis, which is random. To address this challenge, the secondary users are classified into four priority classes by maintaining a separate queue for each priority class. The assignment of channels to secondary users starts from the highest priority queue towards the least priority queue. Further, the QoS to real-time secondary users is improved by introducing timeout-aware inter-queuing (TAIQ), in which shifting of the secondary user is allowed from low-priority queue to high-priority queue when its time deadline is about to reach the time deadlines of its next high level priority queue SUs. The results showed that the proposed four priorities with timeout-aware inter-queuing mechanism give less blocking to real-time secondary users with respect to conventional real-time and non-real-time secondary users.

This paper showcases a prototype system for predicting and visualizing distributed product information queues at one of the largest global appliance companies, BSH Hausgerate GmbH (BSH). A design science methodology is employed to develop a viable artifact (a prediction and visualization system) as a technology-based solution to an important and relevant business problem (product information updates in a distributed environment) using novel solutions (machine learning and visualization), to demonstrate the value of the design (through user interviews), and to disseminate the results to both technical and managerial audiences. The results contribute to the ongoing academic discourse regarding the applications of artificial intelligence to business and the practical development of prediction and visualization solutions for similar environments in terms of distributed environment complexity and company size.

In the paper, we consider a queuing system with n types of customers. We assume that each customer arrives at the queue according to a renewal process and takes a random resource amount, independent of their service time. We write Kolmogorov integro-differential equation, which, in general, cannot be analytically solved. Hence, we look for the solution under the condition of infinitely growing a service time, and we obtain multi-dimensional asymptotic approximations. We show that the n-dimensional probability distribution of the total resource amounts is asymptotically Gaussian, and we look at its accuracy via Kolmogorov distance.

The queueing system considered consist of two multi-server stations in series. Customers arrive according to a Markovian Arrival Process to an infinite capacity queue at the first station. There are c servers who provide identical exponentially distributed service at the first station. A customer at the head of the queue can enter into service if any one of the servers at the first stage is idle. At the second station there are N identical servers called specialist servers . The service time distribution of specialist severs is phase type. There is a finite buffer in between the two stations. On completion of service at first stage, a customer needs service at the second station with probability p or leaves the system with probability 1 − p. In the former case, the customer joins the second station for service in case the waiting room is not full, else he is lost to the system. A customer in the finite buffer can enter into service if at least one of these servers is free. Stability of the system is established and stationary distribution is obtained using Matrix Analytic Methods. We compute distribution of waiting time of customers in the first queue, the mean number of customers lost due to capacity restriction of the waiting space of the second station and the mean waiting time of customers who get into service at the second station. An optimization problem on the capacity of second waiting station is also analyzed.

The discrete-time re-sequencing model, consisting of one high and one low priority finite-capacity queue and a single server, which serves the low priority queue if and only if the high priority queue is empty, is being considered. Two types of customers, regular and re-sequencing, arrive at the system. The arrival and service processes are geometric, i.e. in each time slot at most one customer of each type may arrive at the system and at most one customer may be served. A regular customer upon arrival occupies one place in the high priority queue. An arriving re-sequencing customer moves one customer from the high priority queue (if it is not empty) to the low priority queue and itself leaves the system. A regular customer which sees the high priority queue full and a re-sequenced customer which sees the low priority queue full, are lost. Using the generating function method the recursive procedure for the computation of the joint stationary distribution of the number of customers in the high and in the low priority queues is derived.

Ready queue processing time estimation problem deals with many constraints. Because the processes which reside in the ready queue of computer memory come in varieties such as process size, process requirement indifferences and process types. To match up all these differences is a difficult task to solve so that the processes can be used to perform its task efficiently at any platform. A prior estimation of ready queue processing time helps to meet the system reliability and robustness. A pre-calculated time will ensure the system from failure; also, the backup of task performed can be maintained. In this paper, the existing methods on this approach are described, and how some new methods can be used for the better performance is demonstrated. For this purpose, some sampling techniques are used, and the lottery scheduling procedure is explained which very efficiently performs this task of scheduling on the basis of probabilistic approach and randomness property. The estimation is performed by using sampling methods; with the help of some mathematical calculations, the results are obtained, and finally, confidence interval will ensure the accuracy of the result. So that some new methods can be generated; this provides the result as more efficient than the previous ones. Although various scheduling schemes are available, the lottery scheduling scheme provides the fairness and also removes starvation. Rather than working on the complete data set, some samples can be generated to modularize the work which will be efficient too. So, this paper proposed some new methods in ready queue processing time estimation in multiprocessor environment.

Stabilizing the Quality of Service (QoS) currently being viewed as the main challenge in cloud computing environment and can be achieved either by employing an efficient queuing model or robust scheduling and allocation policies with load balancing features or combination of both. The QoS attributes are analyzed in a parallel queuing model featuring a pure analytical approach and extend enhancement to our previous work. The aim of the paper is to study behavior of hybrid parallel queuing model under conditions of three time varying input parameters: arrival rate, service rate, and number of servers on the response variables: queue length, waiting time, response time, and server utilization factor. With the change in these input parameters, the impact on response variables is analyzed to arrive at a valid conclusion.

This analysis is devoted to model retrial queue with, Bernoulli feedback, preferred as well as impatient customers or units in discrete environment. Here, server follows state-dependent policy and may leave for single vacation whenever it is idle. This investigation is motivated by the increasing impact of discrete-time retrial queues in real-world scenarios. For instance, it is widely used in multiplexing of voice data, digital communication (ATMs, BISDN), switching modules and networks, etc. In such types of queueing system, time is a random variable of discrete type, and we calculate it in equally sized data units. We have considered a system with early arrival pattern and studied the Markov’s chain underlying this model. Along with this, the marginal generating function (mgf) for the total units present in the orbit depends on the server state. Few performance measures like average units in the orbit are also calculated by applying probability generating function method. Further, a practical example is numerically illustrated as well as sensitivity analysis is provided.