R. B. Lenin

Computing packet loss probabilities in multiplexer models using rational approximation

A statistical multiplexer is a basic model used in the design and the dimensioning of communication networks. The multiplexer model consists of a single server queue with constant service time and a more or less complicated arrival process. The aim is to determine the packet loss probability as a function of the capacity of the buffer. In this paper, we show how rational approximation techniques may be applied to compute the packet loss efficiently. The approach is based on the knowledge of a limited number of sample values, together with the decay rate of the probability distribution function. A strategy is proposed where the sample points are chosen automatically. The accuracy of the approach is validated by comparison with both analytical results obtained using a matrix-analytic method and simulation results.

Arrival Time Based Traffic Signal Optimization for Intelligent Transportation Systems

Road Transportation is a crucial component of todays society, which drives several facets of our lives. The goal of intelligent transportation systems (ITS) is to improve the effectiveness, efficiency, and safety of the transportation system. Traffic signals are an elementary component of all road transportation systems. In order to maximize the productivity of a city, traffic signals must be able to efficiently control the flow of vehicles. Traditionally, current traffic signal optimization is based on traffic arrival rates, either estimated or forecasted. In this paper, we illustrate that arrival time based solutions can outperform arrival rate based approaches. To the best of our knowledge, this is the first work that exploits arrival times of vehicles to improve traffic signal efficiency in order to reduce stopped delays and fuel consumptions, thus in turn reducing greenhouse gases and emissions. We show that arrival time knowledge can be utilized in realizing drastic gains in sparse load scenarios and significant gains in moderate load scenarios. The performance improvement translates to reducing stopped delays by over 40,000 hours daily and in reducing fuel consumption by over 650 gallons/signal/day.

Performance analysis of wireless sensor networks using queuing networks

Wireless Sensor Networks (WSNs) are autonomous wireless systems consists of a variety of collaborative sensor nodes forming a self-configuring network with or without any pre-defined infrastructure. The common challenges of a WSN are network connectivity, node mobility, energy consumption, data computation and aggregation at sensor nodes. In this paper we focus on intermittency in network connectivity due to mobility of sensor nodes. We propose a new mathematical model to capture a given entire WSN as is with intermittency introduced between the communication links due to mobility. The model involves open GI/G/1/N queuing networks whereby intermittency durations in communication links are captured in terms of mobility models. The analytical formulas for the performance measures such as average end-to-end delay, packet loss probability, throughput, and average number of hops are derived using the queuing network analyzer and expansion method for models with infinite- and finite-buffer nodes, respectively. For models with 2-state intermittency, we analyze the performance measures by classifying these models into three types: namely, model with intermittent reception, model with intermittent transmission and/or reception, and model with intermittent transmission. We extend the analysis to multi-state intermittency models. We demonstrate the gained insight of WSNs through extensive numerical results.

Differentiated service strategies for ad-hoc wireless sensor networks in harsh communication environments

In some wireless sensor network applications, sensor nodes will be deployed in harsh communication environments. In such environments, the deployment may not be adequately controlled, and nodes may have to communicate with a single destination node. For nodes to alert the destination on critical data that has been sensed, in addition to the harsh communication environment, contention resulting from both the deployment and network density must be appropriately overcome. In this paper, we create theoretical models for the behavior of Timeout-MAC (T-MAC) protocol, and evaluate five possible solutions, each designed to be easy to implement on a device by simply tuning T-MAC parameters, so as to overcome these environment-specific issues and effectively alert the destination to critical data. Our results indicate that slight changes to the behavior of the network can improve the awareness of the destination to critical regions in the environment, and that these changes have different levels of effectiveness at different network densities.

Computing packet loss probabilities of D-BMAP/PH/1/N queues with group services

Statistical multiplexers have been integral components of packet switches and routers on data networks. They are modeled as queueing systems with a finite buffer space, served by one or more transmission links of fixed or varying capacity. The service structure typically admits packets of multiple sources on a first-come first-serve (FCFS) basis. In this paper, we adhere to D-BMAP/PH/1/N queues with discrete phase-type group service channel which allows the packets to get service in the service channel for a random number of time slots by staying in different phases of the service channel before they leave the switch. The aim is to determine the packet loss probability (PLP) as a function of the capacity of the buffer. Due to the curse of dimensionality of the mathematical model, the numerical computation of the performance measures using the analytical formulas is time and memory consuming. Due to rare events, getting the performance measures by simulation is again time consuming. To overcome this problem, we use the Newton-Pade-type rational approximation technique to compute the PLP more efficiently. Since this technique needs the asymptotic behavior of the PLP, we show a way to regroup the elements of the TPM to obtain the asymptotic behavior of the PLP.

Adaptive buffering scheme to reduce packet loss on densely connected WSN with mobile sink

A buffer overflow at a network node occurs if the sum of the amount of data it generates and the amount of data it receives from upstream nodes exceeds its transmission throughput. Densely connected wireless sensor networks (WSNs) with intermittently available sinks (such as mobile sinks) strongly inherit such a buffer overflow problem as sensor nodes in WSNs have limited buffering capacity, and the intermittent nature of these sinks can further influence the network throughput. We propose distribution and caching of sensor-generated data at sensor nodes that belong to any of the minimum cost paths leading to a mobile sink path. We further propose to store the data at each source node for a certain period of time instead of immediately forwarding them to the downstream nodes so as to reduce unnecessary network congestion. Both mathematical and simulation evaluation show that these schemes significantly reduce packet loss probability at network of any size. Consequently, the amount of information lost in the system due to packet overflow/retransmission will be significantly reduced.

A system of systems approach: A benchmark to WSNs mobility models

In last years conference, we presented a system of systems (SoS) approach for the development of a generalized mathematical model for modeling complex systems with integrated WSNs so that necessary benchmark cases can be derived. Additionally, this will allow for a standardized evaluation of performance measures for the different mobility models of WSNs. The generalized model has been tested mathematically and verified be simulation. In this paper, specific mobility models have been tested against this framework and the results have shown that the generalized model provides a powerful framework for the verification of QoS attributes of different models.

Impact of Duty Cycle Variation on WSNs

Wireless Sensor Networks (WSNs) operate with varying duty cycles to meet their application-specific criteria such as the availability, reliability, and the life expectancy of the system. This variation in duty cycle consequently affects the system characteristics including the interference and collision of signals. However, the sensitivity to physical jamming attacks with respect to duty cycle of the network is not widely explored area of research. This paper presents a detailed analysis of the effect of the duty cycle to the interference and collision of signals in the WSNs. In particular, our simulation model depicts a log normal shadowing model to represent a realistic wireless channel and observe the effects of duty cycle variation when a physical jamming attack is launched on the network using a compromised node. Our results show that setting the duty cycle at predetermined value would help minimize the packet drop ratio of the network.

Optimizing appointment template and number of staff of an OB/GYN clinic – micro and macro simulation analyses

BackgroundThe Department of Obstetrics and Gynecology (OB/GYN) at the University of Arkansas for Medical Sciences (UAMS) tested various, new system-restructuring ideas such as varying number of different types of nurses to reduce patient wait times for its outpatient clinic, often with little or no effect on waiting time. Witnessing little progress despite these time-intensive interventions, we sought an alternative way to intervene the clinic without affecting the normal clinic operations.AimThe aim is to identify the optimal (1) time duration between appointments and (2) number of nurses to reduce wait time of patients in the clinic.MethodsWe developed a discrete-event computer simulation model for the OB/GYN clinic. By using the patient tracker (PT) data, appropriate probability distributions of service times of staff were fitted to model different variability in staff service times. These distributions were used to fine-tune the simulation model. We then validated the model by comparing the simulated wait times with the actual wait times calculated from the PT data. The validated model was then used to carry out “what-if” analyses.ResultsThe best scenario yielded 16 min between morning appointments, 19 min between afternoon appointments, and addition of one medical assistant. Besides removing all peak wait times and bottlenecks around noon and late in the afternoon, the best scenario yielded 39.84 % (p<.001), 30.31 % (p<.001), and 15.12 % (p<.001) improvement in patients’ average wait times for providers in the exam rooms, average total wait time at various locations and average total spent time in the clinic, respectively. This is achieved without any compromise in the utilization of the staff and in serving all patients by 5pm.ConclusionsA discrete-event simulation model is developed, validated, and used to carry out “what-if” scenarios to identify the optimal time between appointments and number of nurses. Using the model, we achieved a significant improvement in wait time of patients in the clinic, which the clinic management initially had difficulty achieving through manual interventions. The model provides a tool for the clinic management to test new ideas to improve the performance of other UAMS OB/GYN clinics.

Development of generalized HPC simulator

With ever growing interests in High Performance Computer (HPC) systems for their inherent capabilities for data consolidation, data modeling, systems simulation, data storage, or intensive computation solutions there will obviously be ever growing needs from researchers within governments, businesses, and academia to determine the right HPC framework for meeting their computational needs for the least cost. While HPC simulators are available from commercial sources, they are not free and often commit a group into design choices that are not always optimal. In this short paper, we present our preliminary effort to design an open-source, adaptable, free-standing and extendible HPC simulator.