Browsing by Author "Raissi-Dehkordi, Majid"
Results Per Page
Sort Options
Item Broadcast Scheduling for Push Broadcast Systems with Arbitrary Cost Functions(2007) Raissi-Dehkordi, Majid; Baras, John S.; ISRIn this report the problem of broadcast scheduling in Push broadcast systems is studied. We introduce an optimization approach that leads to well justified policies for Push broadcast systems with generalized cost functions. In particular, we apply our results to a Push broadcast system with different deadlines associated to the files while allowing the files to have unequal demand rates and lengths. We will show that our proposed policy covers some of the previously investigated Push systems as special cases and is applicable to a wide range of cost functions assigned to the files. We also calculate the optimal average cost for our experimental settings and show, through extensive simulation studies, that our results closely match that value for each experiment.Item Broadcast Scheduling in Information Delivery Networks(2002) Raissi-Dehkordi, Majid; Baras, John S.; ISR; CSHCNThe continuous growth in the demand for access to information andthe increasing number of users of the information delivery systemshave sparked the need for highly scalable systems with moreefficient usage of the bandwidth. One of the effective methods forefficient use of the bandwidth is to provide the information to agroup of users simultaneously via broadcast delivery. Generally,all applications that deliver the popular data packages (trafficinformation, weather, stocks, web pages) are suitable candidatesfor broadcast delivery and satellite or wireless networks withtheir inherent broadcast capability are the natural choices forimplementing such applications.In this dissertation, we investigate one of the most importantproblems in broadcast delivery i.e., the broadcast schedulingproblem. This problem arises in broadcast systems with a largenumber of data packages and limited broadcast channels and thegoal is to find the best sequence of broadcasts in order tominimize the average waiting time of the users.
We first formulate the problem as a dynamic optimization problemand investigate the properties of the optimal solution. Later, weuse the bandit problem formulation to address a version of theproblem where all packages have equal lengths. We find anasymptotically optimal index policy for that problem and comparethe results with some well-known heuristic methods.
Since the equal file length assumption is not appropriate forapplications such as cache broadcasting in the Internet deliverysystems, we also investigate an extension of the problem where thefiles have random lengths. After investigating some of theproperties of the optimal solution, we derive an asymptoticallyoptimal index policy for that case as well. Also, throughsimulation studies, the performance of the policy is compared withthat of some other heuristic polices designed by intuitivearguments. The index policy is also extended to systems withdeterministic, unequal file sizes and its performance is evaluatedand compared to other policies via simulation studies.
The formulation and analytical procedures used in deriving theindex policies in this dissertation allow for introduction ofother extensions of the problem like assigning weights to the datafiles (studied in Chapter 3) or taking into account the channelerrors and correlation between the arrivals. We will present ourformulation of the last two extensions and discuss some of thenumerical results to motivate future work on these problems.
Item A Dynamic Optimization Approach to the Scheduling Problem in Satellite and Wireless Broadcast Systems(2002) Raissi-Dehkordi, Majid; Baras, John S.; ISR; CSHCNThe continuous growth in the demand for access to information andthe increasing number of users of the information delivery systemshave sparked the need for highly scalable systems with moreefficient usage of the bandwidth. One of the effective methods forefficient use of the bandwidth is to provide the information to agroup of users simultaneously via broadcast delivery. Generally,all applications that deliver the popular data packages (trafficinformation, weather, stocks, web pages) are suitable candidatesfor broadcast delivery and satellite or wireless networks withtheir inherent broadcast capability are the natural choices forimplementing such applications.In this report, we investigate one of the most important problemsin broadcast delivery i.e., the broadcast scheduling problem. Thisproblem arises in broadcast systems with a large number of datapackages and limited broadcast channels and the goal is to findthe best sequence of broadcasts in order to minimize the average waiting time of the users.We first formulate the problem as a dynamic optimization problemand investigate the properties of the optimal solution. Later, weuse the bandit problem formulation to address a version of theproblem where all packages have equal lengths. We find anasymptotically optimal index policy for that problem and comparethe results with some well-known heuristic methods.Item Hybrid Internet Simulation Testbed(1999) Liu, Mingyan D.; Karir, Manish; Raissi-Dehkordi, Majid; Baras, John S.; Baras, John S.; ISR; CSHCNInternet technology as a widely accepted modern telecommunicationstandard has been widely extended to combine with numerous othertechnologies, e.g., satellite, ATM, wireless. This is whatwe term Hybrid Internet. Along with this technology emerging,various enhancements and alterations of standard TCP/IP for differentpurposes have been proposed and studied intensively. More and morefrequently we are facing the question of how to choose from thesedifferent schemes to design a system for a particular purpose, whichwould inevitably involve the interaction and trade-off study. We believethat simulation is a powerful tool for this type of work.In this paper, we describeour implementation of a Hybrid Internet testbed which includesa series of traffic models and TCP/IP enhancements.The goal of our work is to makea set of reusable modules upon which we can build complex systemsto study the standard protocols and their variations.We also present application examples using these module components.Item Interactive Data Services in Wireless Access Networks: Capacity Planning and Protocols(2005) Zhou, Xiaoming; Raissi-Dehkordi, Majid; Baras, John S.; Baras, John S.; ISR; CSHCNIn this paper, we study the capacity planning in wireless access network for interactive data services such as web browsing. A closed queuing model has been developed which can capture the bottleneck effects in both the forward and the reverse channels. The model can be used to calculate the average throughput, the average response time and the number of users the system can support. We evaluate the performance of several MAC protocols such as slotted Aloha, static TDMA, Aloha/periodic stream and combined free demand assignment multiple access (CFDAMA) using realistic web traffic models. Based on the performance evaluation, we propose a new MAC protocol and a new transport layer protocol. Our new MAC protocol called combined polling free demand assignment multiple access (CPFDAMA) explores the correlation between forward channel data packets and reverse channel acknowledgement packets. Our new transport layer protocol called RWBP uses per-flow queuing, round robin scheduling and receiver window backpressure for congestion management. RWBP can eliminate congestion losses inside the wireless networks. Our protocol suite outperforms the proposed protocols in term of both channel utilization and response time. Our results can be used for service providers to dimension their networks and provide quality of service to a certain number of users.Item UAV Placement for Enhanced Connectivity in wireless Ad-hoc Networks(2004) Raissi-Dehkordi, Majid; Chandrashekar, Karthikeyan; Baras, John S.; ISR; CSHCNIn this paper we address the problem of providing full connectivity in large (wide area) ad hoc networks by placing advantaged nodes like UAVs (as relay nodes) in appropriate places. We provide a formulation where we can treat the connectivity problem as a clustering problem with a summation-form distortion function. We then adapt the Deterministic Annealing clustering algorithm to our formulation and using that we nd the minimum number of UAVs required to provide connectivity and their locations. Furthermore, we describe enhancements that can be used to extend the basic connectivity problem to support notions of reliable connectivity that can lead to improved network performance. We establish the validity of our algorithm and compare its performance with optimal (exhaustive search) as well as non-opitmal (hard clustering) algorithms.We show that our algorithm is nearoptimal both for the basic connectivity problem as well as extended notions of connectivity.