SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
The application of NASA management approach to solving complex problems on earth is discussed. The management of the Apollo program is presented as an example of effective management techniques. Four key elements of effective management are analyzed. Photographs of the Cape Kennedy launch sites and supporting equipment are included to support the discussions.
In this paper we discuss the need for interactive factory schedule repair and improvement, and we identify case-based reasoning (CBR) as an appropriate methodology. Case-based reasoning is the problem solving paradigm that relies on a memory for past problem solving experiences (cases) to guide current problem solving. Cases similar to the current case are retrieved from the case memory, and similarities and differences of the current case to past cases are identified. Then a best case is selected, and its repair plan is adapted to fit the current problem description. If a repair solution fails, an explanation for the failure is stored along with the case in memory, so that the user can avoid repeating similar failures in the future. So far we have identified a number of repair strategies and tactics for factory scheduling and have implemented a part of our approach in a prototype system, called CABINS. As a future work, we are going to scale up CABINS to evaluate its usefulness in a real manufacturing environment.
Critical evaluation of systems analysis techniques for solving social problems
Deep Learning for Enhancing Fault TolerantCapabilities of Scientific WorkflowsAlok Singh, Ilkay AltintasSan Diego Supercomputer Center, UCSDLa Jolla, CA, USA{a1singh, ialtintas}@ucsd.eduMalachi Schram, Nathan TallentPacific Northwestern National LaboratoryRichland, WA, USA{malachi.schram, nathan.tallent}@pnnl.govAbstract-In the history of Computer Science, the act of ‘delegation’ has been the greatest multiplier of society’s problem-solving ability. A scientist working on detecting anomalies in a phenomenon, does not need to re-invent matrix multiplication techniques to solve her problem. Scientific workflows provide ultimate ‘delegation’ mechanism -where a domain scientist can completely forget the specifics of ‘how’ her program will execute on a large cluster in an efficient and cost-effective manner and can instead focus on the mathematical formulation and theoretical robustness of her solution. We present here an approach that directly aims to make the execution of Scientific Workflows more reliable, robust and efficient. We aim that the work presented in this paper will propel the larger effort, from the scientific workflow community, of making scientific workflow execution as simple, efficient and robust as a JOIN operation in a modern database. Specifically, we apply Deep Learning techniques to develop a mechanism that forecasts the final state (success or failure) of a dynamic job in a large-scale particle physics experiment, with minimal data gathering, and as early as possible in job’s life cycle. The key advantage of having a predictive mechanism to identify and anticipate failure-prone jobs is the potential for designing intelligent Fault Tolerance mechanisms to handle anomalous events. We achieve a 14% improvement in computational resources utilization, and an overall classification accuracy of 85% on real tasks executed in a High Energy Physics Computing workflow.To the best of our knowledge, this is the most exhaustive and first of its kind study of neural network architectures in context of a real-dataset profiled from a large-scalescientific workflow.
The potential achievements of solar system exploration are outlined, and a course of action is suggested which will maximize the rewards. Also provided is a sourcebook of information on the solar system and the technology being brought to bear for its exploration. The document explores the degree to which three practical questions can be answered: why it is necessary to explore the solar system, why understanding of the solar system is important to us, and why we cannot wait until all terrestrial problems are solved before an attempt is made to solve problems in space.
Research in knowledge representation, machine learning, and knowledge acquisition performed at Knowledge Systems Lab. is summarized. The major goal of the research was to develop flexible, effective methods for representing the qualitative knowledge necessary for solving large problems that require symbolic reasoning as well as numerical computation. The research focused on integrating different representation methods to describe different kinds of knowledge more effectively than any one method can alone. In particular, emphasis was placed on representing and using spatial information about three dimensional objects and constraints on the arrangement of these objects in space. Another major theme is the development of robust machine learning programs that can be integrated with a variety of intelligent systems. To achieve this goal, learning methods were designed, implemented and experimented within several different problem solving environments.
In this paper, the use of hybrid expert system shells and hybrid (i.e., algorithmic and heuristic) approaches for solving engineering problems is reported. Aspects of various engineering problem domains are reviewed for a number of examples with specific applications made to recently developed prototype expert systems. Based on this prototyping experience, critical evaluations of and comparisons between commercially available tools, and some research tools, in the United States and Australia, and their underlying problem-solving paradigms are made. Characteristics of the implementation tool and the engineering domain are compared and practical software engineering issues are discussed with respect to hybrid tools and approaches. Finally, guidelines are offered with the hope that expert system development will be less time consuming, more effective, and more cost-effective than it has been in the past.
The use of experimental and testbed vehicles and their contributions to fundamental problem solving and overall vehicle characteristics are presented. A chronological description of aircraft from the first supersonic flight in the X-1 through the development of the F-100 series, the X-15, and the B-1 is given. One of the early significant contributions made was the development of the all-moving stabilizer that solved the problem associated with high-speed tuck and the trim requirements of entry into the supersonic regime. Some wind tunnel/flight drag characteristics studies associated with the B-1 bomber involving flexibility effects and wind tunnel effects are shown. Finally, the evolution of digital systems to enhance maintainability and reliability and to reduce the work load in the cockpit are described.
PVEX is described as an expert system that solves the problem of selection of the material and process in missile manufacturing. The producibility and the value problem has been deeply studied in the past years, and was written in dBase III and PROLOG before. A new approach is presented in that the solution is achieved by introducing hypothetical reasoning, heuristic criteria integrated with a simple hypertext system and shell programming. PVEX combines KMS with Unix scripts which graphically depicts decision trees. The decision trees convey high level qualitative problem solving knowledge to users, and a stand-alone help facility and technical documentation is available through KMS. The system developed is considerably less development costly than any other comparable expert system.
Both crew productivity and flight safety depend critically on efficient crew problem solving and decision making. But what constitutes efficient problem solving and decision making behavior? Previous work has focused on crew performance in the context of a specific problem type. The present study differs from these analyses in two respects. First, we describe how aircraft crews cope with different types of problems. And second, we attempt to discern characteristics of optimal task management behavior. For this reason, we contrasted the behavior of more effective and less effective crews. Underlying this approach is the assumption that crews whom commit few operational errors also demonstrate most efficient problem solving and decision making behavior.
Head-up display (HUD) unit mounted in welder's visor or helmet provides process information in real time. Display enables welder to gain greater awareness of effects of welding while controlling welding process in usual way(s) by use of foot pedal, voice command, and/or manipulation of welding torch. Possible to avoid need for every new welder on job to solve same problems solved by other welders. Graphical display presented to welder, adjacent to normal field of view through visor. Beneficial in supporting and enhancing continued practice of manual arc welding. Valuable in industries that invest millions of technician-hours each year in manual arc welding for variety of products.
An evolutionary algorithm applies evolution-based principles to problem solving. To solve a problem, the user defines the space of potential solutions, the representation space. Sample solutions are encoded in a chromosome-like structure. The algorithm maintains a population of such samples, which undergo simulated evolution by means of mutation, crossover, and survival of the fittest principles. Genetic Programming (GP) uses tree-like chromosomes, providing very rich representation suitable for many problems of interest. GP has been successfully applied to a number of practical problems such as learning Boolean functions and designing hardware circuits. To apply GP to a problem, the user needs to define the actual representation space, by defining the atomic functions and terminals labeling the actual trees. The sufficiency principle requires that the label set be sufficient to build the desired solution trees. The closure principle allows the labels to mix in any arity-consistent manner. To satisfy both principles, the user is often forced to provide a large label set, with ad hoc interpretations or penalties to deal with undesired local contexts. This unfortunately enlarges the actual representation space, and thus usually slows down the search. In the past few years, three different methodologies have been proposed to allow the user to alleviate the closure principle by providing means to define, and to process, constraints on mixing the labels in the trees. Last summer we proposed a new methodology to further alleviate the problem by discovering local heuristics for building quality solution trees. A pilot system was implemented last summer and tested throughout the year. This summer we have implemented a new revision, and produced a User's Manual so that the pilot system can be made available to other practitioners and researchers. We have also designed, and partly implemented, a larger system capable of dealing with much more powerful heuristics.
Robotic mining could prove to be an efficient method of mining resources for extended missions on the Moon or Mars. One component of robotic mining is scouting an area for resources to be mined by other robotic systems. Writing controllers for scouting can be difficult due to the need for fault tolerance, inter-agent cooperation, and agent problem solving. Reinforcement learning could solve these problems by enabling the scouts to learn to improve their performance over time. This work is divided into two sections, with each section addressing the use of machine learning in this domain. The first contribution of this work focuses on the application of reinforcement learning to mining mission analysis. Various mission parameters were modified and control policies were learned. Then agent performance was used to assess the effect of the mission parameters on the performance of the mission. The second contribution of this work explores the potential use of reinforcement learning to learn a controller for the scouts. Through learning, these scouts would improve their ability to map their surroundings over time.
Modification of Potter method of Gaussian elimination for solving eigenvalue problems of buckling and free vibrations of shells of revolution
In this study, the program named FANS-3D (Finite Analytic Numerical Simulation-3 Dimensional) is presented. FANS-3D was designed to solve problems of incompressible fluid flow and combined modes of heat transfer. It solves problems with conduction and convection modes of heat transfer in laminar flow, with provisions for radiation and turbulent flows. It can solve singular or conjugate modes of heat transfer. It also solves problems in natural convection, using the Boussinesq approximation. FANS-3D was designed to solve heat transfer problems inside one, two and three dimensional geometries that can be represented by orthogonal planes in a Cartesian coordinate system. It can solve internal and external flows using appropriate boundary conditions such as symmetric, periodic and user specified.
Data accessibility and transfer of information, including the land resources information system pilot, are structured as large computer information networks. These pilot efforts include the reduction of the difficulty to find and use data, reducing processing costs, and minimize incompatibility between data sources. Artificial Intelligence (AI) techniques were suggested to achieve these goals. The applicability of certain AI techniques are explored in the context of distributed problem solving systems and the pilot land data system (PLDS). The topics discussed include: PLDS and its data processing requirements, expert systems and PLDS, distributed problem solving systems, AI problem solving paradigms, query processing, and distributed data bases.
Pseudospectral methods are a highly accurate means of solving continuous problems. They suffer the Gibbs’ phenomenon in presence of discontinuities. The Gibbs phenomenon can be robustly removed in post-processing via edge detection and mollification (arXiv:1712.09952, Piotrowska et al. in prep.) Taking advantage of the robustness of mollification, we are currently extending our working framework to 2D.
A program of research in the field of artificial intelligence is presented. The research areas discussed include automatic theorem proving, representations of real-world environments, problem-solving methods, the design of a programming system for problem-solving research, techniques for general scene analysis based upon television data, and the problems of assembling an integrated robot system. Major accomplishments include the development of a new problem-solving system that uses both formal logical inference and informal heuristic methods, the development of a method of automatic learning by generalization, and the design of the overall structure of a new complete robot system. Eight appendices to the report contain extensive technical details of the work described.