DESIRE P is a general purpose continuous time simulation language suitable for interactive simulation, dynamic system study, mathematical modeling, process control analysis. It includes an interactive editor, file manipulation facilities, and graphic packages, making it a completely self-contained system. The PDP-11 version of DESIRE P handles 20 state variables, while the VAX/VMS version runs 150 or more. An interpreted job-control language serves for interactive program entry, editing and file operations, and for programming multirun simulation studies. The dynamic segment, containing differential equations in first-order form, is entered just like the job-control statments and accesses the same variables. DESIRE P is largely written in PASCAL, and most of it can be transferred to different computers, with little change. The PASCAL implementation proves that the high-level language can be used to program direct executing languages, still keeping efficiency and speed comparable to assembly language. The runtime compiler of DESIRE P generates fast and efficient code. DESIRE P can incorporate existing and new precompiled FORTRAN numerical integration algorithms.

The phenomenal growth of Information Technology requires us to elicit, store and maintain huge volumes of data. Analyzing this data for various purposes is becoming increasingly important. Data mining consists of applying data analysis and discovery algorithms that under acceptable computational efficiency limitations, produce a particular enumeration of patterns over the data. We present two techniques based on using Logic programming tools for data mining. Data mining analyzes data by extracting patterns which describe its structure and discovers co-relations in the form of rules. We distinguish analysis methods as visual and non-visual and present one application of each. We explain that our focus on the field of Logic Programming makes some of the very complex tasks related to Web based data mining and dynamic content generation, simple and easy to implement in a uniform framework.

The Defense Language Institute (DLI) offers 23 beginning language courses and in 2004 began to provide a smaller class size for these courses. Restrictions on when classes can begin and a limited number of instructors prevent all students from being trained in a smaller class. This thesis develops integer linear programs (ILPs) that generate schedules for all student classes and maximize the number of smaller class starts for a given number of instructors. Secondary scheduling goals include avoiding weekly changes to instructor levels and scheduling preferences such as the number of classes to start simultaneously. The ILPs solve in less than one minute and offer a significant improvement in the number of students that may be trained in the smaller class size. Computational results using real data for the Arabic, Chinese-Mandarin, and Persian-Farsi courses verify the ILPs find feasible multiyear schedules that incorporate the DLI's scheduling preferences while exceeding the DLI's published schedule results. For example, the ILPs find schedules for Arabic that train 8%, 34% and 76% of students in the smaller class in 2006, 2007, and 2008, whereas DLI's manual schedules at best can train 8%, 7% and 64%.

by Shiu-Tung Ng. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1998. / Includes bibliographical references (leaves 100-103). / Abstract also in Chinese. / Chapter 1 --- Preliminary --- p.5 / Chapter 1.1 --- Linear and Convex Optimization Model --- p.5 / Chapter 1.2 --- Notations for Linear Optimization --- p.5 / Chapter 1.3 --- Definition and Properties of Convexities --- p.7 / Chapter 1.4 --- Useful Theorem for Unconstrained Minimization --- p.10 / Chapter 2 --- Linear Optimization --- p.11 / Chapter 2.1 --- Self-dual Linear Optimization Model --- p.11 / Chapter 2.2 --- Definitions and Main Theorems --- p.14 / Chapter 2.3 --- Self-dual Embedding and Simple Example --- p.22 / Chapter 2.4 --- Newton step --- p.25 / Chapter 2.5 --- "Rescaling and Definition of δ(xs,w)" --- p.29 / Chapter 2.6 --- An Interior Point Method --- p.32 / Chapter 2.6.1 --- Algorithm with Full Newton Steps --- p.33 / Chapter 2.6.2 --- Iteration Bound --- p.33 / Chapter 2.7 --- Background and Rounding Procedure for Interior-point Solution --- p.36 / Chapter 2.8 --- Solving Some LP problems --- p.42 / Chapter 2.9 --- Remarks --- p.51 / Chapter 3 --- Convex Optimization --- p.53 / Chapter 3.1 --- Introduction --- p.53 / Chapter 3.1.1 --- Convex Optimization Problem --- p.53 / Chapter 3.1.2 --- Idea of Interior Point Method --- p.55 / Chapter 3.2 --- Logarithmic Barrier Method --- p.55 / Chapter 3.2.1 --- Basic Concepts and Properties --- p.55 / Chapter 3.2.2 --- k-Self-Concordance Condition --- p.62 / Chapter 3.2.3 --- Short-step Logarithmic Barrier Algorithm --- p.64 / Chapter 3.2.4 --- Initialization Algorithm --- p.67 / Chapter 3.3 --- Center Method --- p.70 / Chapter 3.3.1 --- Basic Concepts and Properties --- p.70 / Chapter 3.3.2 --- Short-step Center Algorithm --- p.75 / Chapter 3.3.3 --- Initialization Algorithm --- p.76 / Chapter 3.4 --- Properties and Examples on Self-Concordance --- p.78 / Chapter 3.5 --- Examples of Convex Optimization Problem --- p.82 / Chapter 3.5.1 --- Self-concordant Logarithmic Barrier and Distance Function --- p.82 / Chapter 3.5.2 --- General Convex Optimization Problems --- p.91 / Chapter 3.6 --- Remarks --- p.98 / Bibliography

Providing robots with the ability to grasp objects has, despite decades of research, remained a challenging problem. The problem is approachable in constrained environments where there is ample prior knowledge of the scene and objects that will be manipulated. The challenge is in building systems that scale beyond specific situational instances and gracefully operate in novel conditions. In the past, heuristic and simple rule based strategies were used to accomplish tasks such as scene segmentation or reasoning about occlusion. These heuristic strategies work in constrained environments where a roboticist can make simplifying assumptions about everything from the geometries of the objects to be interacted with, level of clutter, camera position, lighting, and a myriad of other relevant variables. With these assumptions in place, it becomes tractable for a roboticist to hardcode desired behaviour and build a robotic system capable of completing repetitive tasks. These hardcoded behaviours will quickly fail if the assumptions about the environment are invalidated. In this thesis we will demonstrate how a robust grasping system can be built that is capable of operating under a more variable set of conditions without requiring significant engineering of behavior by a roboticist. This robustness is enabled by a new found ability to empower novel machine learning techniques with massive amounts of synthetic training data. The ability of simulators to create realistic sensory data enables the generation of massive corpora of labeled training data for various grasping related tasks. The use of simulation allows for the creation of a wide variety of environments and experiences exposing the robotic system to a large number of scenarios before ever operating in the real world. This thesis demonstrates that it is now possible to build systems that work in the real world trained using deep learning on synthetic data. The sheer volume of data that can be produced via simulation enables the use of powerful deep learning techniques whose performance scales with the amount of data available. This thesis will explore how deep learning and other techniques can be used to encode these massive datasets for efficient runtime use. The ability to train and test on synthetic data allows for quick iterative development of new perception, planning and grasp execution algorithms that work in a large number of environments. Creative applications of machine learning and massive synthetic datasets are allowing robotic systems to learn skills, and move beyond repetitive hardcoded tasks.

Ng, Chi Kong. / "May 2003." / Thesis (Ph.D.)--Chinese University of Hong Kong, 2003. / Includes bibliographical references (p. 175-187). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. Ann Arbor, MI : ProQuest Information and Learning Company, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Mode of access: World Wide Web. / Abstracts in English and Chinese.

Digitized by Kansas Correctional Industries

Programming by demonstration is an end-user programming technique that allows people to create programs by showing the computer examples of what they want to do. Users do not need specialised programming skills. Instead, they instruct the computer by demonstrating examples, much as they might show another person how to do the task. Programming by demonstration empowers users to create programs that perform tedious and time-consuming computer chores. However, it is not in widespread use, and is instead confined to research applications that end users never see. This makes it difficult to evaluate programming by demonstration tools and techniques. This thesis claims that domain-independent programming by demonstration can be made available in existing applications and used to automate iterative tasks by end users. It is supported by Familiar, a domain-independent, AppleScript-based programming-by-demonstration tool embodying standard machine learning algorithms. Familiar is designed for end users, so works in the existing applications that they regularly use. The assertion that programming by demonstration can be made available in existing applications is validated by identifying the relevant platform requirements and a range of platforms that meet them. A detailed scrutiny of AppleScript highlights problems with the architecture and with many implementations, and yields a set of guidelines for designing applications that support programming-by-demonstration. An evaluation shows that end users are capable of using programming by demonstration to automate iterative tasks. However, the subjects tended to prefer other tools, choosing Familiar only when the alternatives were unsuitable or unavailable. Familiar's inferencing is evaluated on an extensive set of examples, highlighting the tasks it can perform and the functionality it requires.

We develop a framework for finding robust solutions of constraint programs. Our approach is based on the notion of fault tolerance. We formalise this concept within constraint programming, extend it in several dimensions and introduce some algorithms to find robust solutions efficiently. When applying constraint programming to real world problems we often face uncertainty. Whilst reactive methods merely deal with the consequences of an unexpected change, taking a more proactive approach may guarantee a certain level of robustness. We propose to apply the fault tolerance framework, introduced in [Ginsberg 98], to constraint programming: A robust solution is one such that a small perturbation only requires a small response. We identify, define and classify a number of abstract problems related to stability within constraint satisfaction or optimisation. We propose some efficient and effective algorithms for solving these problems. We then extend this framework by allowing the repairs and perturbations themselves to be constrained. Finally, we assess the practicality of this framework on constraint satisfaction and scheduling problems.

Thesis (M.S.)--Rochester Institute of Technology, 2007. / Typescript. Includes bibliographical references (leaves 74-78).