Efficient Content-based Routing, Mobility-aware Topologies, and Temporal Subspace Matching

Event-based systems are seen as good candidates for supporting distributed applications in dynamic and ubiquitous environments because they support decoupled and asynchronous many-to-many information dissemination. Event systems are widely used, because asynchronous messaging provides a flexible alternative to RPC (Remote Procedure Call). They are typically implemented using an overlay network of routers. A content-based router forwards event messages based on filters that are installed by subscribers and other routers. The filters are organized into a routing table in order to forward incoming events to proper subscribers and neighbouring routers. This thesis addresses the optimization of content-based routing tables organized using the covering relation and presents novel data structures and configurations for improving local and distributed operation. Data structures are needed for organizing filters into a routing table that supports efficient matching and runtime operation. We present novel results on dynamic filter merging and the integration of filter merging with content-based routing tables. In addition, the thesis examines the cost of client mobility using different protocols and routing topologies. We also present a new matching technique called temporal subspace matching. The technique combines two new features. The first feature, temporal operation, supports notifications, or content profiles, that persist in time. The second feature, subspace matching, allows more expressive semantics, because notifications may contain intervals and be defined as subspaces of the content space. We also present an application of temporal subspace matching pertaining to metadata-based continuous collection and object tracking.

Hybrid Simulation of Queuing-System Dynamics

This paper presents a simple hybrid computer technique to study the transient behaviour of queueing systems. This method is superior to stand-alone analog or digital solution because the hardware requirement is excessive for analog technique whereas computation time is appreciable in the latter case. By using a hybrid computer one can share the analog hardware thus requiring fewer integrators. The digital processor can store the values, play them back at required time instants and change the coefficients of differential equations. By speeding up the integration on the analog computer it is feasible to solve a large number of these equations very fast. Hybrid simulation is even superior to the analytic technique because in the latter case it is difficult to solve time-varying differential equations.

Computationally Efficient Methods for MDL-Optimal Density Estimation and Data Clustering

The Minimum Description Length (MDL) principle is a general, well-founded theoretical formalization of statistical modeling. The most important notion of MDL is the stochastic complexity, which can be interpreted as the shortest description length of a given sample of data relative to a model class. The exact definition of the stochastic complexity has gone through several evolutionary steps. The latest instantation is based on the so-called Normalized Maximum Likelihood (NML) distribution which has been shown to possess several important theoretical properties. However, the applications of this modern version of the MDL have been quite rare because of computational complexity problems, i.e., for discrete data, the definition of NML involves an exponential sum, and in the case of continuous data, a multi-dimensional integral usually infeasible to evaluate or even approximate accurately. In this doctoral dissertation, we present mathematical techniques for computing NML efficiently for some model families involving discrete data. We also show how these techniques can be used to apply MDL in two practical applications: histogram density estimation and clustering of multi-dimensional data.

Efficient search for statistically significant dependency rules in binary data

Analyzing statistical dependencies is a fundamental problem in all empirical science. Dependencies help us understand causes and effects, create new scientific theories, and invent cures to problems. Nowadays, large amounts of data is available, but efficient computational tools for analyzing the data are missing. In this research, we develop efficient algorithms for a commonly occurring search problem - searching for the statistically most significant dependency rules in binary data. We consider dependency rules of the form X->A or X->not A, where X is a set of positive-valued attributes and A is a single attribute. Such rules describe which factors either increase or decrease the probability of the consequent A. A classical example are genetic and environmental factors, which can either cause or prevent a disease. The emphasis in this research is that the discovered dependencies should be genuine - i.e. they should also hold in future data. This is an important distinction from the traditional association rules, which - in spite of their name and a similar appearance to dependency rules - do not necessarily represent statistical dependencies at all or represent only spurious connections, which occur by chance. Therefore, the principal objective is to search for the rules with statistical significance measures. Another important objective is to search for only non-redundant rules, which express the real causes of dependence, without any occasional extra factors. The extra factors do not add any new information on the dependence, but can only blur it and make it less accurate in future data. The problem is computationally very demanding, because the number of all possible rules increases exponentially with the number of attributes. In addition, neither the statistical dependency nor the statistical significance are monotonic properties, which means that the traditional pruning techniques do not work. As a solution, we first derive the mathematical basis for pruning the search space with any well-behaving statistical significance measures. The mathematical theory is complemented by a new algorithmic invention, which enables an efficient search without any heuristic restrictions. The resulting algorithm can be used to search for both positive and negative dependencies with any commonly used statistical measures, like Fisher's exact test, the chi-squared measure, mutual information, and z scores. According to our experiments, the algorithm is well-scalable, especially with Fisher's exact test. It can easily handle even the densest data sets with 10000-20000 attributes. Still, the results are globally optimal, which is a remarkable improvement over the existing solutions. In practice, this means that the user does not have to worry whether the dependencies hold in future data or if the data still contains better, but undiscovered dependencies.

Modeling Efficient Classification as a Process of Confidence Assessment and Delegation

In visual object detection and recognition, classifiers have two interesting characteristics: accuracy and speed. Accuracy depends on the complexity of the image features and classifier decision surfaces. Speed depends on the hardware and the computational effort required to use the features and decision surfaces. When attempts to increase accuracy lead to increases in complexity and effort, it is necessary to ask how much are we willing to pay for increased accuracy. For example, if increased computational effort implies quickly diminishing returns in accuracy, then those designing inexpensive surveillance applications cannot aim for maximum accuracy at any cost. It becomes necessary to find trade-offs between accuracy and effort. We study efficient classification of images depicting real-world objects and scenes. Classification is efficient when a classifier can be controlled so that the desired trade-off between accuracy and effort (speed) is achieved and unnecessary computations are avoided on a per input basis. A framework is proposed for understanding and modeling efficient classification of images. Classification is modeled as a tree-like process. In designing the framework, it is important to recognize what is essential and to avoid structures that are narrow in applicability. Earlier frameworks are lacking in this regard. The overall contribution is two-fold. First, the framework is presented, subjected to experiments, and shown to be satisfactory. Second, certain unconventional approaches are experimented with. This allows the separation of the essential from the conventional. To determine if the framework is satisfactory, three categories of questions are identified: trade-off optimization, classifier tree organization, and rules for delegation and confidence modeling. Questions and problems related to each category are addressed and empirical results are presented. For example, related to trade-off optimization, we address the problem of computational bottlenecks that limit the range of trade-offs. We also ask if accuracy versus effort trade-offs can be controlled after training. For another example, regarding classifier tree organization, we first consider the task of organizing a tree in a problem-specific manner. We then ask if problem-specific organization is necessary.

An improved simulation model to predict pre-harvest aflatoxin risk in maize

Aflatoxin is a potent carcinogen produced by Aspergillus flavus, which frequently contaminates maize (Zea mays L.) in the field between 40° north and 40° south latitudes. A mechanistic model to predict risk of pre-harvest contamination could assist in management of this very harmful mycotoxin. In this study we describe an aflatoxin risk prediction model which is integrated with the Agricultural Production Systems Simulator (APSIM) modelling framework. The model computes a temperature function for A. flavus growth and aflatoxin production using a set of three cardinal temperatures determined in the laboratory using culture medium and intact grains. These cardinal temperatures were 11.5 °C as base, 32.5 °C as optimum and 42.5 °C as maximum. The model used a low (≤0.2) crop water supply to demand ratio—an index of drought during the grain filling stage to simulate maize crop's susceptibility to A. flavus growth and aflatoxin production. When this low threshold of the index was reached the model converted the temperature function into an aflatoxin risk index (ARI) to represent the risk of aflatoxin contamination. The model was applied to simulate ARI for two commercial maize hybrids, H513 and H614D, grown in five multi-location field trials in Kenya using site specific agronomy, weather and soil parameters. The observed mean aflatoxin contamination in these trials varied from <1 to 7143 ppb. ARI simulated by the model explained 99% of the variation (p ≤ 0.001) in a linear relationship with the mean observed aflatoxin contamination. The strong relationship between ARI and aflatoxin contamination suggests that the model could be applied to map risk prone areas and to monitor in-season risk for genotypes and soils parameterized for APSIM.

An improved simulation model to predict pre-harvest aflatoxin risk in maize

Aflatoxin is a potent carcinogen produced by Aspergillus flavus, which frequently contaminates maize (Zea mays L.) in the field between 40° north and 40° south latitudes. A mechanistic model to predict risk of pre-harvest contamination could assist in management of this very harmful mycotoxin. In this study we describe an aflatoxin risk prediction model which is integrated with the Agricultural Production Systems Simulator (APSIM) modelling framework. The model computes a temperature function for A. flavus growth and aflatoxin production using a set of three cardinal temperatures determined in the laboratory using culture medium and intact grains. These cardinal temperatures were 11.5 °C as base, 32.5 °C as optimum and 42.5 °C as maximum. The model used a low (≤0.2) crop water supply to demand ratio—an index of drought during the grain filling stage to simulate maize crop's susceptibility to A. flavus growth and aflatoxin production. When this low threshold of the index was reached the model converted the temperature function into an aflatoxin risk index (ARI) to represent the risk of aflatoxin contamination. The model was applied to simulate ARI for two commercial maize hybrids, H513 and H614D, grown in five multi-location field trials in Kenya using site specific agronomy, weather and soil parameters. The observed mean aflatoxin contamination in these trials varied from <1 to 7143 ppb. ARI simulated by the model explained 99% of the variation (p ≤ 0.001) in a linear relationship with the mean observed aflatoxin contamination. The strong relationship between ARI and aflatoxin contamination suggests that the model could be applied to map risk prone areas and to monitor in-season risk for genotypes and soils parameterized for APSIM.

Electronic Building Passport - Final Report: Project 2 National Energy Efficient Building Project Phase 2

The National Energy Efficient Building Project (NEEBP) Phase One report, published in December 2014, investigated “process issues and systemic failures” in the administration of the energy performance requirements in the National Construction Code. It found that most stakeholders believed that under-compliance with these requirements is widespread across Australia, with similar issues being reported in all states and territories. The report found that many different factors were contributing to this outcome and, as a result, many recommendations were offered that together would be expected to remedy the systemic issues reported. To follow up on this Phase 1 report, three additional projects were commissioned as part of Phase 2 of the overall NEEBP project. This Report deals with the development and piloting of an Electronic Building Passport (EBP) tool – a project undertaken jointly by pitt&sherry and a team at the Queensland University of Technology (QUT) led by Dr Wendy Miller. The other Phase 2 projects cover audits of Class 1 buildings and issues relating to building alterations and additions. The passport concept aims to provide all stakeholders with (controlled) access to the key documentation and information that they need to verify the energy performance of buildings. This trial project deals with residential buildings but in principle could apply to any building type. Nine councils were recruited to help develop and test a pilot electronic building passport tool. The participation of these councils – across all states – enabled an assessment of the extent to which these councils are currently utilising documentation; to track the compliance of residential buildings with the energy performance requirements in the National Construction Code (NCC). Overall we found that none of the participating councils are currently compiling all of the energy performance-related documentation that would demonstrate code compliance. The key reasons for this include: a major lack of clarity on precisely what documentation should be collected; cost and budget pressures; low public/stakeholder demand for the documentation; and a pragmatic judgement that non-compliance with any regulated documentation requirements represents a relatively low risk for them. Some councils reported producing documentation, such as certificates of final completion, only on demand, for example. Only three of the nine council participants reported regularly conducting compliance assessments or audits utilising this documentation and/or inspections. Overall we formed the view that documentation and information tracking processes operating within the building standards and compliance system are not working to assure compliance with the Code’s energy performance requirements. In other words the Code, and its implementation under state and territory regulatory processes, is falling short as a ‘quality assurance’ system for consumers. As a result it is likely that the new housing stock is under-performing relative to policy expectations, consuming unnecessary amounts of energy, imposing unnecessarily high energy bills on occupants, and generating unnecessary greenhouse gas emissions. At the same time, Councils noted that the demand for documentation relating to building energy performance was low. All the participant councils in the EBP pilot agreed that documentation and information processes need to work more effectively if the potential regulatory and market drivers towards energy efficient homes are to be harnessed. These findings are fully consistent with the Phase 1 NEEBP report. It was also agreed that an EBP system could potentially play an important role in improving documentation and information processes. However, only one of the participant councils indicated that they might adopt such a system on a voluntary basis. The majority felt that such a system would only be taken up if it were: - A nationally agreed system, imposed as a mandatory requirement under state or national regulation; - Capable of being used by multiple parties including councils, private certifiers, building regulators, builders and energy assessors in particular; and - Fully integrated into their existing document management systems, or at least seamlessly compatible rather than a separate, unlinked tool. Further, we note that the value of an EBP in capturing statistical information relating to the energy performance of buildings would be much greater if an EBP were adopted on a nationally consistent basis. Councils were clear that a key impediment to the take up of an EBP system is that they are facing very considerable budget and staffing challenges. They report that they are often unable to meet all community demands from the resources available to them. Therefore they are unlikely to provide resources to support the roll out of an EBP system on a voluntary basis. Overall, we conclude from this pilot that the public good would be well served if the Australian, state and territory governments continued to develop and implement an Electronic Building Passport system in a cost-efficient and effective manner. This development should occur with detailed input from building regulators, the Australian Building Codes Board (ABCB), councils and private certifiers in the first instance. This report provides a suite of recommendations (Section 7.2) designed to advance the development and guide the implementation of a national EBP system.

Monte Carlo simulation of Ising chains showing several phase transitions

A Monte Carlo simulation of Ising chains with competing short-range and infiniterange interactions has been carried out. Results show that whenever the system does not enter a metastable state, variation of temperature brings about phase transitions in the Ising chain. These phase transitions, except for two sets of interaction strengths, are generally of higher order and involve changes in the long-range order while the short-range order remains unaffected.

Mathematical modelling and numerical simulation of phosphine flow during grain fumigation in leaky cylindrical silos

The phosphine distribution in a cylindrical silo containing grain is predicted. A three-dimensional mathematical model, which accounts for multicomponent gas phase transport and the sorption of phosphine into the grain kernel is developed. In addition, a simple model is presented to describe the death of insects within the grain as a function of their exposure to phosphine gas. The proposed model is solved using the commercially available computational fluid dynamics (CFD) software, FLUENT, together with our own C code to customize the solver in order to incorporate the models for sorption and insect extinction. Two types of fumigation delivery are studied, namely, fan- forced from the base of the silo and tablet from the top of the silo. An analysis of the predicted phosphine distribution shows that during fan forced fumigation, the position of the leaky area is very important to the development of the gas flow field and the phosphine distribution in the silo. If the leak is in the lower section of the silo, insects that exist near the top of the silo may not be eradicated. However, the position of a leak does not affect phosphine distribution during tablet fumigation. For such fumigation in a typical silo configuration, phosphine concentrations remain low near the base of the silo. Furthermore, we find that half-life pressure test readings are not an indicator of phosphine distribution during tablet fumigation.

A heterogeneous computing approach to simulation of the Heston Stochastic Volatility Model

Stochastic volatility models are of fundamental importance to the pricing of derivatives. One of the most commonly used models of stochastic volatility is the Heston Model in which the price and volatility of an asset evolve as a pair of coupled stochastic differential equations. The computation of asset prices and volatilities involves the simulation of many sample trajectories with conditioning. The problem is treated using the method of particle filtering. While the simulation of a shower of particles is computationally expensive, each particle behaves independently making such simulations ideal for massively parallel heterogeneous computing platforms. In this paper, we present our portable Opencl implementation of the Heston model and discuss its performance and efficiency characteristics on a range of architectures including Intel cpus, Nvidia gpus, and Intel Many-Integrated-Core (mic) accelerators.