Hongya Wang

Scheduling Periodic Continuous Queries in Real-Time Data Broadcast Environments

On-demand broadcast is a promising data dissemination approach in mobile computing environments thanks to its adaptability and scalability for large-scale and dynamic workload. An important class of emerging data broadcast applications needs to monitor multiple time-varying data items continuously to be kept aware of the up-to-date information. This paper investigates the broadcast schedule problem for disseminating timely data to periodic continuous queries, and a systematic and highly efficient solution for applications of this type is provided. In particular, we propose a novel measure, called Bandwidth Utilization, to quantify the minimum bandwidth demand of a periodic continuous query set. The timing predictability can be ensured if a set of periodic continuous queries passes a bandwidth utilization based schedulability test. The schedulability test techniques are also extended to deal with dynamic query arrival and departure. An efficient online scheduling algorithm, called RM-UO, is developed, which can fulfill the timing constraints combined with the proposed query release and deletion policies. To demonstrate the effectiveness of theoretical results, an illustrative algorithm implementation is presented along with comprehensive performance analysis. Simulation results show that our solution offers nice timing predictability whereas other comparable best effort scheduling algorithms such as SIN-Q and DTIU experience different deadline miss ratios at different query workloads.

On a novel property of the earliest deadline first algorithm

Real-time scheduling theory plays a key role in many time critical control systems or applications. In this paper, an interesting property of the Earliest Deadline First (EDF) algorithm, which has never been discussed before, is examined. To be specific, we conjecture that if a task set is schedulable under EDF, then for any task pair (τi, τj) such that pi ≥ pj in this task set, there must be at least one whole execution of τj occurring between the release time and deadline of any τis job. Although this property is not hard to describe, its proof is far more difficult than expected. To prove this property, we first show the correctness of the conjecture for task sets consisting of only two real-time tasks. In view of the hardness in extending the proof to task sets having more than 2 members, extensive simulation experiments are conducted to support our intuition for general cases. The conjecture holds under a substantial number of parameter settings we have tried.

Locality sensitive hashing revisited: filling the gap between theory and algorithm analysis

Locality Sensitive Hashing (LSH) is widely recognized as one of the most promising approaches to similarity search in high-dimensional spaces. Based on LSH, a considerable number of nearest neighbor search algorithms have been proposed in the past, with some of them having been used in many real-life applications. Apart from their demonstrated superior performance in practice, the popularity of the LSH algorithms is mainly due to their provable performance bounds on query cost, space consumption and failure probability. In this paper, we show that a surprising gap exists between the LSH theory and widely practiced algorithm analysis techniques. In particular, we discover that a critical assumption made in the classical LSH algorithm analysis does not hold in practice, which suggests that using the existing methods to analyze the performance of practical LSH algorithms is a conceptual mismatch. To address this problem, a novel analysis model is developed that bridges the gap between the LSH theory and the method for analyzing the LSH algorithm performance. With the help of this model, we identify some important flaws in the commonly used analysis methods in the LSH literature. The validity of this model is verified through extensive experiments with real datasets.

A time-sensitive personalized recommendation method based on probabilistic matrix factorization technique

Personalized recommender systems are the most effective way to solve the problem of information overload. The majority of traditional personalized recommender systems employ the collaborative filtering (CF) approach. CF leverages users’ behaviors to infer a target user’s preference for a particular item, while ignores the fact that users interact with the system within a particular context, such as a particular time interval or location. In this paper, we propose a novel time-sensitive personalized recommendation method called TSPR for movie recommendation. Specifically, we first define and construct a new user–context rating matrix based on the original user–movie rating matrix and then propose a novel context-dependent similarity measurement by mining the implicit relationship among users from the user–context rating matrix. Further, we build a context-dependent similarity matrix based on the context-dependent similarity measurement. Finally, we incorporate the context-dependent similarity matrix into the probabilistic matrix factorization model. The experimental results show that TSPR performs much better than the state-of-the-art recommendation methods.

Hyperbolic Utilization Bounds for Rate Monotonic Scheduling on Homogeneous Multiprocessors

The utilization bounds for partitioned multiprocessor scheduling are a function of task allocation algorithms and the schedulability conditions selected for uniprocessor scheduling algorithms. In this paper, we use rate-monotonic scheduling on each processor and present the lower and upper limits of the utilization bounds for all reasonable task allocation heuristics. Unlike previous work, the hyperbolic bound due to Bini , instead of the Liu & Layland bound, is adopted to do the schedulability test on uniprocessors. We also derive the utilization bounds with respect to the worst fit allocation algorithm and reasonable allocation decreasing heuristics, and the two bounds are found to coincide with the worst and best achievable multiprocessor utilization bounds, respectively. Analytical and experimental results show that the proposed utilization bound performs better than the existing bound under quite a lot of parameter settings, and combining these two bounds together can significantly (up to 3 times) increase the number of schedulable task sets with little extra overhead.

The Hyperbolic Schedulability Bound for Multiprocessor RM Scheduling

To verify the feasibility of real-time task sets on homogeneous multiprocessor systems, Lopez et al. derived the utilization bound based on Best Fit Decreasing allocation algorithm and Rate-Monotonic scheduling, which coincides with the maximum achievable multiprocessor utilization bound, using Liu & Lyland bound as the uniprocessor schedulability condition. In this paper, a novel feasibility test for the same target problem is developed based on the hyperbolic bound due to Bini et al., instead of the Liu & Layland bound. Analytical and experimental results show that the proposed utilization bound performs better than the existing bound under quite a lot of parameter settings, and combining these two bounds together can significantly increase the number of schedulable task sets with little extra overhead.

Why locality sensitive hashing works: A practical perspective

Abstract Locality Sensitive Hashing (LSH) is one of the most efficient approaches to the nearest neighbor search problem in high dimensional spaces. A family H of hash functions is called locality sensitive if the collision probability p h ( r ) of any two points 〈 q , p 〉 at distance r over a random hash function h decreases with r. The classic LSH algorithm employs a data structure consisting of k ⁎ l randomly chosen hash functions to achieve more desirable collision curves and the collision probability P h k l ( r ) for 〈 q , p 〉 is equal to 1 − ( 1 − p h ( r ) k ) l . The great success of LSH is usually attributed to the solid theoretical guarantee for P h k l ( r ) and p h ( r ) . In practice, however, users are more interested in recall rate, i.e., the probability that a random query collides with its r-near neighbor over a fixed LSH data structure h l k . Implicitly or explicitly, P h k l ( r ) is often misinterpreted as recall rate and used to predict the performance of LSH. This is problematic because P h k l ( r ) is actually the expectation of recall rates. Interestingly, numerous empirical studies show that, for most (if not all) real datasets and a fixed sample of random LSH data structure, the recall rate is very close to P h k l ( r ) . In this paper, we provide a theoretical justification for this phenomenon. We show that (1) for random datasets the recall rate is asymptotically equal to P h k l ( r ) ; (2) for arbitrary datasets the variance of the recall rate is very small as long as the parameter k and l are properly chosen and the size of datasets is large enough. Our analysis (1) explains why the practical performance of LSH (the recall rate) matches so well with the theoretical expectation ( P h k l ( r ) ); and (2) indicates that, in addition to the nice theoretical guarantee, the mechanism by which LSH data structures are constructed and the huge amount of data are also the main causes for the success of LSH in practice.

Towards load shedding and scheduling schemes for data streams that maintain quality and timing requirements of query results

Real-time stream processing is essential for many real-life stream-based applications. Systems designed to run such applications must be prepared to operate under overloaded conditions. In this paper, the load shedding problem is studied for an important class of real-time data stream monitoring applications. In particular, we adopt the

Incorporating Contextual Information into a Mobile Advertisement Recommender System

(m,k)