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.
Uniform tasks distributed uniformly or else assigned to one processor. Investigation in theory of scheduling yielded optimal scheme for allocation of tasks among digital data processors in hypercube ensemble. Applies to tasks that require equal time to execute, performed in any order, and between any two of which equal amounts of communication required. Reduces overall processing time for given set of computational tasks.
Researchers explore an approach to using general purpose parallel computers which involves mapping hardware resources onto computations instead of mapping computations onto hardware. Problems such as processor allocation, task scheduling and load balancing, which have traditionally proven to be challenging, change significantly under this approach and may become amenable to new attacks. Researchers describe the implementation of this approach used by the FFP Machine whose computation and communication resources are repeatedly partitioned into disjoint groups that match the needs of available tasks from moment to moment. Several consequences of this system are examined.
The current task of Resource Allocation Process includes the planning and apportionment of JPL's Ground Data System composed of the Deep Space Network and Mission Control and Computing Center facilities. The addition of the data driven, rule based planning system, RALPH, has expanded the planning horizon from 8 weeks to 10 years and has resulted in large labor savings. Use of the system has also resulted in important improvements in science return through enhanced resource utilization. In addition, RALPH has been instrumental in supporting rapid turn around for an increased volume of special what if studies. The status of RALPH is briefly reviewed and important lessons learned from the creation of an highly functional design team are focused on through an evolutionary design and implementation period in which an AI shell was selected, prototyped, and ultimately abandoned, and through the fundamental changes to the very process that spawned the tool kit. Principal topics include proper integration of software tools within the planning environment, transition from prototype to delivered to delivered software, changes in the planning methodology as a result of evolving software capabilities and creation of the ability to develop and process generic requirements to allow planning flexibility.
The preliminary results of a research program investigating the use of added tasks to evaluate mental workload are reviewed. The focus of the first studies was a reappraisal of the traditional secondary task logic that encouraged the use of low-priority instructions for the added task. It was believed that such low-priority tasks would encourage subjects to split their available resources among the two tasks. The primary task would be assigned all the resources it needed, and any remaining reserve capacity would be assigned to the secondary task. If the model were correct, this approach was expected to combine sensitivity to primary task difficulty with unintrusiveness to primary task performance. The first studies of the current project demonstrated that a high-priority added task, although intrusive, could be more sensitive than the traditional low-priority secondary task. These results suggested that a more appropriate model of the attentional effects associated with added task performance might be based on capacity switching, rather than the traditional optimal allocation model.
The concept of constraint propagation was discussed. Performance increases are possible with careful application of these constraint mechanisms. The degree of performance increase is related to the interdependence of the different activities resource usage. Although this method of applying constraints to activities and resources is often beneficial, it is obvious that this is no panacea cure for the computational woes that are experienced by dynamic resource allocation and scheduling problems. A combined effort for execution optimization in all areas of the system during development and the selection of the appropriate development environment is still the best method of producing an efficient system.
An assessment of the planned processing hardware and software/firmware for the Communications and Tracking System of the Space Station Freedom (SSF) was performed. The intent of the assessment was to determine the optimum distribution of software/firmware in the processing hardware for maximum throughput with minimum required memory. As a product of the assessment process an assessment methodology was to be developed that could be used for similar assessments of future manned spacecraft system designs. The assessment process was hampered by changing requirements for the Space Station. As a result, the initial objective of determining the optimum software/firmware allocation was not fulfilled, but several useful conclusions and recommendations resulted from the assessment. It was concluded that the assessment process would not be completely successful for a system with changing requirements. It was also concluded that memory requirements and hardware requirements were being modified to fit as a consequence of the change process, and although throughput could not be quantitized, potential problem areas could be identified. Finally, inherent flexibility of the system design was essential for the success of a system design with changing requirements. Recommendations resulting from the assessment included development of common software for some embedded controller functions, reduction of embedded processor requirements by hardwiring some Orbital Replacement Units (ORUs) to make better use of processor capabilities, and improvement in communications between software development personnel to enhance the integration process. Lastly, a critical observation was made regarding the software integration tasks did not appear to be addressed in the design process to the degree necessary for successful satisfaction of the system requirements.
This invention relates to a resource allocation in computer systems, and more particularly, to a method and associated apparatus for shortening response time and improving efficiency of a heterogeneous distributed networked computer system by reallocating the jobs queued up for busy nodes to idle, or less-busy nodes. In accordance with the algorithm (SIDA for short), the load-sharing is initiated by the server device in a manner such that extra overhead in not imposed on the system during heavily-loaded conditions. The algorithm employed in the present invention uses a dual-mode, server-initiated approach. Jobs are transferred from heavily burdened nodes (i.e., over a high threshold limit) to low burdened nodes at the initiation of the receiving node when: (1) a job finishes at a node which is burdened below a pre-established threshold level, or (2) a node is idle for a period of time as established by a wakeup timer at the node. The invention uses a combination of the local queue length and the local service rate ratio at each node as the workload indicator.
The following topics are presented in view graph form: the background of the Resource Allocation Planning Helper (RALPH); RALPH schedule lifecycle; scheduling approach; technology layering; details; and RALPH directions.
A VLSI implementable neural network architecture for dynamic assignment is presented. The resource allocation problems involve assigning members of one set (e.g. resources) to those of another (e.g. consumers) such that the global 'cost' of the associations is minimized. The network consists of a matrix of sigmoidal processing elements (neurons), where the rows of the matrix represent resources and columns represent consumers. Unlike previous neural implementations, however, association costs are applied directly to the neurons, reducing connectivity of the network to VLSI-compatible 0 (number of neurons). Each row (and column) has an additional neuron associated with it to independently oversee activations of all the neurons in each row (and each column), providing a programmable 'k-winner-take-all' function. This function simultaneously enforces blocking (excitatory/inhibitory) constraints during convergence to control the number of active elements in each row and column within desired boundary conditions. Simulations show that the network, when implemented in fully parallel VLSI hardware, offers optimal (or near-optimal) solutions within only a fraction of a millisecond, for problems up to 128 resources and 128 consumers, orders of magnitude faster than conventional computing or heuristic search methods.
Several buffer structures are discussed and compared in terms of implementation complexity, interswitch handshaking requirements, and their ability to deal with variations in traffic patterns and message lengths. A new design of buffers is presented that provide non-FIFO message handling and efficient storage allocation for variable size packets using linked lists managed by a simple on-chip controller. The new buffer design is evaluated by comparing it to several alternative designs in the context of a multistage interconnection network. The present modeling and simulations show that the new buffer outperforms alternative buffers and can thus be used to improve the performance of a wide variety of systems currently using less efficient buffers.
An overview of the 1992 World Administrative Radio Conference is presented with viewgraphs. Allocations of radio frequency spectrum are addressed. Mobile satellite service, broadcast satellite service, and uplink power control beacons are also addressed.
The problem of routing jobs to K parallel queues with identical exponential servers and unequal finite buffer capacities is considered. Routing decisions are taken by a controller which has buffering space available to it and may delay routing of a customer to a queue. Using ideas from weak majorization, it is shown that the shorter nonfull queue delayed (SNQD) policy minimizes both the total number of customers in the system at any time and the number of customers that are rejected by that time. The SNQD policy always delays routing decisions as long as all servers are busy. Only when all the buffers at the controller are occupied is a customer routed to the queue with the shortest queue length that is not at capacity. Moreover, it is shown that, if a fixed number of buffers is to be distributed among the K queues, then the optimal allocation scheme is the one in which the difference between the maximum and minimum queue capacities is minimized, i.e., becomes either 0 or 1.
Rocket engine design follows three phases: systems design, parameter design, and tolerance design. Systems design and parameter design are most effectively conducted in a concurrent engineering (CE) environment that utilize methods such as Quality Function Deployment and Taguchi methods. However, tolerance allocation remains an art driven by experience, handbooks, and rules of thumb. It was desirable to develop and optimization approach to tolerancing. The case study engine was the STME gas generator cycle. The design of the major components had been completed and the functional relationship between the component tolerances and system performance had been computed using the Generic Power Balance model. The system performance nominals (thrust, MR, and Isp) and tolerances were already specified, as were an initial set of component tolerances. However, the question was whether there existed an optimal combination of tolerances that would result in the minimum cost without any degradation in system performance.
Traditionally, allocation of data in distributed database management systems has been determined by off-line analysis and optimization. This technique works well for static database access patterns, but is often inadequate for frequently changing workloads. In this paper we address how to dynamically reallocate data for partionable distributed databases with changing access patterns. Rather than complicated and expensive optimization algorithms, a simple heuristic is presented and shown, via an implementation study, to improve system throughput by 30 percent in a local area network based system. Based on artificial wide area network delays, we show that dynamic reallocation can improve system throughput by a factor of two and a half for wide area networks. We also show that individual site load must be taken into consideration when reallocating data, and provide a simple policy that incorporates load in the reallocation decision.
In previous reports, including one by the author, learning has been shown to benefit by having discriminanda move rather than remain stationary. This stimulus movement effect might be attributed to several theoretical mechanisms, including attention, topological memory, and exposure duration. The series of experiments reported in this article was designed to Contrast these potential explanatory factors. Ten rhesus monkeys (Macaca mulatta) were tested on a variety of computerized tasks in which the stimuli remained stationary, flashed, or moved at systematically varied speeds. Performance was significantly best when the sample stimulus moved quickly and was poorest when the stimulus remained stationary. Further analysis of these data and other previously published data revealed that the distribution of the stimulus movement effect across trials supported an attention allocation interpretation.
Robust passification is considered for uncertain linear, time invariant, systems having redundant actuators and sensors. The approach is to obtain optimal sensor blending and control allocation matrices that maximize the region in the parameter space in which the system remains passive. The approach results in a generalized eigenvalue problem consisting of a number of linear matrix inequalities (LMIs). Reduction of the number of LMIs is investigated, and a numerical example is given for demonstrating the approach.
Robust passification is considered for uncertain linear, time invariant systems having redundant actuators and sensors. The approach is to obtain optimal sensor blending and control allocation matrices that maximize the region in the parameter space in which the system remains passive. The approach results in a generalized eigenvalue problem consisting of a number of linear matrix inequalities (LMIs). Reduction of the number of LMIs is investigated, and a numerical example is given for demonstrating the approach.
Design, fabrication and test of a power allocation and distribution (PAD) network for microwave driven actuators is presented in this paper. Development of a circuit that would collect power from a rectenna array amplify and distribute the power to actuators was designed and fabricated for space application in an actuator array driven by a microwave. A P-SPICE model was constructed initially for data reduction purposes, and was followed by a working real-world model. A voltage up - converter (VUC) is used to amplify the voltage from the individual rectenna. The testing yielded a 26:1 voltage amplification ratio with input voltage at 9 volts and a measured output voltage 230VDC. Future work includes the miniaturization of the circuitry, the use of microwave remote control, and voltage amplification technology for each voltage source. The objective of this work is to develop a model system that will collect DC voltage from an array of rectenna and propagate the voltage to an array of actuators.