SEARCH · Engineering Papers

Results for “Maximization”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,081 records · Page 60

All sources NASA NTRS DOE OSTI

NASA NTRS1991

Implementation of a three degree of freedom, motor/brake hybrid force output device for virtual environment control tasks

The advent of high resolution, physical model based computer graphics has left a gap in the design of input/output technology appropriate for interacting with such complex virtual world models. Since virtual worlds consist of physical models, it is appropriate to output the inherent force information necessary for the simulation to the user. The detailed design, construction, and control of a three degree freedom force output joystick will be presented. A novel kinematic design allows all three axes to be uncoupled, so that the system inertia matrix is diagonal. The two planar axes are actuated through an offset gimbal, and the third through a sleeved cable. To compensate for friction and inertia effects, this transmission is controlled by a force feedforward and a closed force feedback proportional loop. Workspace volume is a cone of 512 cubic inches, and the device bandwidth is maximized at 60 Hz for the two planar and 30 Hz for the third axis. Each axis is controlled by a motor/proportional magnetic particle brake combination fixed to the base. The innovative use of motors and brakes allows objects with high resistive torque requirements to be simulated without the stability and related safety issues involved with high torque, energy storing motors alone. Position, velocity, and applied endpoint force are sensed directly. Different control strategies are discussed and implemented, with an emphasis on how virtual environment force information, generated by the MIT Media Lab Computer Graphics and Animation Group BOLIO system, is transmitted to the device controller. The design constraints for a kinesthetic force feedback device can be summarized as: How can the symbiosis between the sense of presence in the virtual environment be maximized without compromising the interaction task under the constraints of the mechanical device limitations? Research in this field will yield insights to the optimal human sensory feedback mix for a wide spectrum of control and interaction problems. A flexible research tool that is designed as an easily reproducible product prototype has been constructed to explore the variety of possible force interaction.

Russo, Massimo↗

NASA NTRS1994

Optimization of Adaptive Intraply Hybrid Fiber Composites with Reliability Considerations

The reliability with bounded distribution parameters (mean, standard deviation) was maximized and the reliability-based cost was minimized for adaptive intra-ply hybrid fiber composites by using a probabilistic method. The probabilistic method accounts for all naturally occurring uncertainties including those in constituent material properties, fabrication variables, structure geometry, and control-related parameters. Probabilistic sensitivity factors were computed and used in the optimization procedures. For actuated change in the angle of attack of an airfoil-like composite shell structure with an adaptive torque plate, the reliability was maximized to 0.9999 probability, with constraints on the mean and standard deviation of the actuation material volume ratio (percentage of actuation composite material in a ply) and the actuation strain coefficient. The reliability-based cost was minimized for an airfoil-like composite shell structure with an adaptive skin and a mean actuation material volume ratio as the design parameter. At a O.9-mean actuation material volume ratio, the minimum cost was obtained.

Shiao, Michael C.↗

NASA NTRS1994

EDOS operations concept and development approach

The Earth Observing System (EOS) Data and Operations System (EDOS) is being developed by the National Aeronautics and Space Administration (NASA) Goddard Space Flight Center (GSFC) for the capture, level zero processing, distribution, and backup archiving of high speed telemetry data received from EOS spacecraft. All data received will conform to the Consultative Committee for Space Data Standards (CCSDS) recommendations. The major EDOS goals are to: (1) minimize EOS program costs to implement and operate EDOS; (2) respond effectively to EOS growth requirements; and (3) maintain compatibility with existing and enhanced versions of NASA institutional systems required to support EOS spacecraft. In order to meet these goals, the following objectives have been defined for EDOS: (1) standardize EDOS interfaces to maximize utility for future requirements; (2) emphasize life-cycle cost (LCC) considerations (rather than procurement costs) in making design decisions and meeting reliability, maintainability, availability (RMA) and upgradability requirements; (3) implement data-driven operations to the maximum extent possible to minimize staffing requirements and to maximize system responsiveness; (4) provide a system capable of simultaneously supporting multiple spacecraft, each in different phases of their life-cycles; (5) provide for technology insertion features to accommodate growth and future LCC reductions during the operations phase; and (6) provide a system that is sufficiently robust to accommodate incremental performance upgrades while supporting operations. Operations concept working group meetings were facilitated to help develop the EDOS operations concept. This provided a cohesive concept that met with approval of responsible personnel from the start. This approach not only speeded up the development process by reducing review cycles, it also provided a medium for generating good ideas that were immediately molded into feasible concepts. The operations concept was then used as a basis for the EDOS specification. When it was felt that concept elements did not support detailed requirements, the facilitator process was used to resolve discrepancies or to add new concept elements to support the specification. This method provided an ongoing revisal of the operations concept and prevented large revisions at the end of the requirement analysis phase of system development.

Knoble, G.↗

NASA NTRS1995

Minimizing structural vibrations with Input Shaping (TM)

A new method for commanding machines to move with increased dynamic performance was developed. This method is an enhanced version of input shaping, a patented vibration suppression algorithm. This technique intercepts a command input to a system command that moves the mechanical system with increased performance and reduced residual vibration. This document describes many advanced methods for generating highly optimized shaping sequences which are tuned to particular systems. The shaping sequence is important because it determines the trade off between move/settle time of the system and the insensitivity of the input shaping algorithm to variations or uncertainties in the machine which can be controlled. For example, a system with a 5 Hz resonance that takes 1 second to settle can be improved to settle instantaneously using a 0.2 shaping sequence (thus improving settle time by a factor of 5). This system could vary by plus or minus 15% in its natural frequency and still have no apparent vibration. However, the same system shaped with a 0.3 second shaping sequence could tolerate plus or minus 40% or more variation in natural frequency. This document describes how to generate sequences that maximize performance, sequences that maximize insensitivity, and sequences that trade off between the two. Several software tools are documented and included.

Singhose, Bill↗

NASA NTRS1994

Effect of spaceflight on skeletal muscle: Mechanical properties and myosin isoform content of a slow muscle

This study examined changes in contractile, biochemical, and histochemical properties of slow antigravity skeletal muscle after a 6-day spaceflight mission. Twelve male Sprague-Dawley rats were randomly divided into two groups: flight and ground-based control. Approximately 3 h after the landing, in situ contractile measurements were made on the soleus muscles of the flight animals. The control animals were studied 24 h later. The contractile measurements included force-velocity relationship, force-frequency relationship, and fatigability. Biochemical measurements focused on the myosin heavy chain (MHC) and myosin light chain profiles. Adenosinetriphosphatase histochemistry was performed to identify cross-sectional area of slow and fast muscle fibers and to determine the percent fiber type distribution. The force-velocity relationships of the flight muscles were altered such that maximal isometric tension P(sub o) was decreased by 24% and maximal shortening velocity was increased by 14% (P less than 0.05). The force-frequency relationship of the flight muscles was shifted to the right of the control muscles. At the end of the 2-min fatigue test, the flight muscles generated only 34% of P(sub o), whereas the control muscles generated 64% of P(sub o). The flight muscles exhibited de novo expression of the type IIx MHC isoform as well as a slight decrease in the slow type I and fast type IIa MHC isoforms. Histochemical analyses of flight muscles demonstrated a small increase in the percentage of fast type II fibers and a greater atrophy of the slow type I fibers. The results demonstrate that contractile properties of slow antigravity skeletal muscle are sensitive to the microgravity environment and that changes begin to occur within the 1st wk. These changes were at least, in part, associated with changes in the amount and type of contractile protein expressed.

Caiozzo, Vincent J.↗

NASA NTRS1994

On the maximum energy release in flux-rope models of eruptive flares

We determine the photospheric boundary conditions which maximize the magnetic energy released by a loss of ideal-magnetohydrodynamic (MHD) equilibrium in two-dimensional flux-rope models. In these models a loss of equilibrium causes a transition of the flux rope to a lower magnetic energy state at a higher altitude. During the transition a vertical current sheet forms below the flux rope, and reconnection in this current sheet releases additional energy. Here we compute how much energy is released by the loss of equilibrium relative to the total energy release. When the flux-rope radius is small compared to its height, it is possible to obtain general solutions of the Grad-Shafranov equation for a wide range of boundary conditions. Variational principles can then be used to find the particular boundary condition which maximizes the magnetic energy released for a given class of conditions. We apply this procedure to a class of models known as cusp-type catastrophes, and we find that the maximum energy released by the loss of equilibrium is 20.8% of the total energy release for any model in this class. If the additional restriction is imposed that the photospheric magnetic field forms a simple arcade in the absence of coronal currents, then the maximum energy release reduces to 8.6%

Forbes, T. G.↗

NASA NTRS1995

Particle chaos and pitch angle scattering

Pitch angle scattering is a factor that helps determine the dawn-to-dusk current, controls particle energization, and it has also been used as a remote probe of the current sheet structure. Previous studies have interpreted their results under the exception that randomization will be greatest when the ratio of the two timescales of motion (gyration parallel to and perpendicular to the current sheet) is closet to one. Recently, the average expotential divergence rate (AEDR) has been calculated for particle motion in a hyperbolic current sheet (Chen, 1992). It is claimed that this AEDR measures the degree of chaos and therefore may be thought to measure the randomization. In contrast to previous expectations, the AEDR is not maximized when Kappa is approximately equal to 1 but instead increases with decreasing Kappa. Also contrary to previous expectations, the AEDR is dependent upon the parameter b(sub z). In response to the challenge to previous expectations that has been raised by this calculation of the AEDR, we have investigated the dependence of a measure of particle pitch angle scattering on both the parameters Kappa and b(sub z). We find that, as was previously expected, particle pitch angle scattering is maximized near Kappa = 1 provided that Kappa/b(sub z) greater than 1. In the opposite regime, Kappa/b(sub z) less than 1, we find that particle pitch angle scattering is still largest when the two timescales are equal, but the ratio of the timescales is proportional to b(sub z). In this second regime, particle pitch angle scattering is not due to randomization, but is instead due to a systematic pitch angle change. This result shows that particle pitch angle scattering need not be due to randomization and indicates how a measure of pitch angle scattering can exhibit a different behavior than a measure of chaos.

Burkhart, G. R.↗

NASA NTRS1995

The distribution of cosmological gamma-ray bursts

We compare the burst distribution of the Burst and Transient Source Experiment (BATSE)-2B catalog to a cosmological distribution. The observed distribution agrees well with a cosmological one, however, it is insensitive to cosmological parameters such as omega and lambda. The bursts are not necessarily standard candles, and their luminosity can vary by up to a factor of 10. The maximal redshift, z(sub max), of bursts longer than 2 s is 2.1(sup +1)(sub -0.7) (assuming no evolution). The present data is insufficient to determine maximal redshift, z(sub max), of bursts shorter than 2 s.

Cohen, Ehud↗

NASA NTRS1994

Estimation and classification by sigmoids based on mutual information

An estimate of the probability density function of a random vector is obtained by maximizing the mutual information between the input and the output of a feedforward network of sigmoidal units with respect to the input weights. Classification problems can be solved by selecting the class associated with the maximal estimated density. Newton's s method, applied to an estimated density, yields a recursive maximum likelihood estimator, consisting of a single internal layer of sigmoids, for a random variable or a random sequence. Applications to the diamond classification and to the prediction of a sun-spot process are demonstrated.

Baram, Yoram↗

NASA NTRS1996

Achieving High Throughput for Data Transfer over ATM Networks

File-transfer rates for ftp are often reported to be relatively slow, compared to the raw bandwidth available in emerging gigabit networks. While a major bottleneck is disk I/O, protocol issues impact performance as well. Ftp was developed and optimized for use over the TCP/IP protocol stack of the Internet. However, TCP has been shown to run inefficiently over ATM. In an effort to maximize network throughput, data-transfer protocols can be developed to run over UDP or directly over IP, rather than over TCP. If error-free transmission is required, techniques for achieving reliable transmission can be included as part of the transfer protocol. However, selected image-processing applications can tolerate a low level of errors in images that are transmitted over a network. In this paper we report on experimental work to develop a high-throughput protocol for unreliable data transfer over ATM networks. We attempt to maximize throughput by keeping the communications pipe full, but still keep packet loss under five percent. We use the Bay Area Gigabit Network Testbed as our experimental platform.

Johnson, Marjory J.↗

NASA NTRS1994

Understanding software faults and their role in software reliability modeling

This study is a direct result of an on-going project to model the reliability of a large real-time control avionics system. In previous modeling efforts with this system, hardware reliability models were applied in modeling the reliability behavior of this system. In an attempt to enhance the performance of the adapted reliability models, certain software attributes were introduced in these models to control for differences between programs and also sequential executions of the same program. As the basic nature of the software attributes that affect software reliability become better understood in the modeling process, this information begins to have important implications on the software development process. A significant problem arises when raw attribute measures are to be used in statistical models as predictors, for example, of measures of software quality. This is because many of the metrics are highly correlated. Consider the two attributes: lines of code, LOC, and number of program statements, Stmts. In this case, it is quite obvious that a program with a high value of LOC probably will also have a relatively high value of Stmts. In the case of low level languages, such as assembly language programs, there might be a one-to-one relationship between the statement count and the lines of code. When there is a complete absence of linear relationship among the metrics, they are said to be orthogonal or uncorrelated. Usually the lack of orthogonality is not serious enough to affect a statistical analysis. However, for the purposes of some statistical analysis such as multiple regression, the software metrics are so strongly interrelated that the regression results may be ambiguous and possibly even misleading. Typically, it is difficult to estimate the unique effects of individual software metrics in the regression equation. The estimated values of the coefficients are very sensitive to slight changes in the data and to the addition or deletion of variables in the regression equation. Since most of the existing metrics have common elements and are linear combinations of these common elements, it seems reasonable to investigate the structure of the underlying common factors or components that make up the raw metrics. The technique we have chosen to use to explore this structure is a procedure called principal components analysis. Principal components analysis is a decomposition technique that may be used to detect and analyze collinearity in software metrics. When confronted with a large number of metrics measuring a single construct, it may be desirable to represent the set by some smaller number of variables that convey all, or most, of the information in the original set. Principal components are linear transformations of a set of random variables that summarize the information contained in the variables. The transformations are chosen so that the first component accounts for the maximal amount of variation of the measures of any possible linear transform; the second component accounts for the maximal amount of residual variation; and so on. The principal components are constructed so that they represent transformed scores on dimensions that are orthogonal. Through the use of principal components analysis, it is possible to have a set of highly related software attributes mapped into a small number of uncorrelated attribute domains. This definitively solves the problem of multi-collinearity in subsequent regression analysis. There are many software metrics in the literature, but principal component analysis reveals that there are few distinct sources of variation, i.e. dimensions, in this set of metrics. It would appear perfectly reasonable to characterize the measurable attributes of a program with a simple function of a small number of orthogonal metrics each of which represents a distinct software attribute domain.

Munson, John C.↗

NASA NTRS1996

Drops in Space: Super Oscillations and Surfactant Studies

An unprecedented microgravity observation of maximal shape oscillations of a surfactant-bearing water drop the size of a ping pong ball was observed during a mission of Space Shuttle Columbia as part of the second United States Microgravity Laboratory-USML-2 (STS-73, October 20-November 5, 1995). The observation was precipitated by the action of an intense sound field which produced a deforming force on the drop. When this deforming force was suddenly reduced, the drop executed nearly free and axisymmetric oscillations for several cycles, demonstrating a remarkable amplitude of nonlinear motion. Whether arising from the discussion of modes of oscillation of the atomic nucleus, or the explosion of stars, or how rain forms, the complex processes influencing the motion, fission, and coalescence of drops have fascinated scientists for centuries. Therefore, the axisymmetric oscillations of a maximally deformed liquid drop are noteworthy, not only for their scientific value but also for their aesthetic character. Scientists from Yale University, the Jet Propulsion Laboratory (JPL) and Vanderbilt University conducted liquid drop experiments in microgravity using the acoustic positioning/manipulation environment of the Drop Physics Module (DPM). The Yale/JPL group's objectives were to study the rheological properties of liquid drop surfaces on which are adsorbed surfactant molecules, and to infer surface properties such as surface tension, Gibb's elasticity, and surface dilatational viscosity by using a theory which relies on spherical symmetry to solve the momentum and mass transport equations.

Apfel, Robert E.↗

NASA NTRS1997

Near-Optimal Re-Entry Trajectories for Reusable Launch Vehicles

A near-optimal guidance law for the descent trajectory for earth orbit re-entry of a fully reusable single-stage-to-orbit pure rocket launch vehicle is derived. A methodology is developed to investigate using both bank angle and altitude as control variables and selecting parameters that maximize various performance functions. The method is based on the energy-state model of the aircraft equations of motion. The major task of this paper is to obtain optimal re-entry trajectories under a variety of performance goals: minimum time, minimum surface temperature, minimum heating, and maximum heading change; four classes of trajectories were investigated: no banking, optimal left turn banking, optimal right turn banking, and optimal bank chattering. The cost function is in general a weighted sum of all performance goals. In particular, the trade-off between minimizing heat load into the vehicle and maximizing cross range distance is investigated. The results show that the optimization methodology can be used to derive a wide variety of near-optimal trajectories.

Chou, H.-C.↗

NASA NTRS1997

Multiple Hub Network Choice in the Liberalized European Market

A key question that so far has received relatively little attention in the germane literature is that of the changes at various airports as a result of the EU liberalization policies. That is, presently, most major European airports still benefit from the so-called home-carrier phenomenon where the country's publicly or semi-publicly owned carrier uses the country's main airport as its gateway hub and, consequently, the home-carrier is also the principal user of this airport (in terms of proportion of total aircraft movements, number of passengers transported, connections, slots ownership, etc.). The country's main airport has substantially benefited from these monopoly conditions of airline captivity, strongly determined by the bilateral system of international air transport regulation. Therefore, European major airports were used to operate in essentially different markets, compared to the increasingly competitive markets of their home based carriers. This partly explains relative stability of transport volumes and financial results of European major airports compared to the relatively volatile financial results of most European national airlines. However, the liberalization of European aviation is likely to change this situation. Market access is open now to all community carriers, i.e. carriers with majority ownership and effective control in the hands of EU citizens. Ticket prices are free, governments can only intervene in case of dumping or excessive pricing. A community airline can choose its seat in any of the 15 member states. Licensing procedures are harmonized between member states. In the last few months community carriers have had unrestricted route access within the EU. Most probably this development will be extended to countries inside and outside Europe. Last year the European Commission got the mandate to start negotiations with 10 other European countries. In the meantime the EC has also started negotiations with the USA on so-called soft rights. In the meantime, open skies agreements have been concluded between the USA and most of the EU member states to facilitate strategic alliances between airlines of the states involved. As a result of this on-going liberalization the model of the single 'national' carrier using the national home base as its single hub for the designated third, fourth and sixth freedom operations will stepwise disappear. Within the EU the concept of the national carrier has already been replaced by that of the community carrier. State ownership in more and more European carriers is reduced. On the longer run mergers or even bankruptcy will further undermine the "single national carrier - single national hub" model in Europe. In the meantime, strategic alliances between national carriers in Europe will already reduce the airlines' loyalty to a single airport. Profit maximization and accountability to share holders will supersede the loyalty of these newly emerging alliances, probably looking for the opportunities of a multiple hub network to adequately cover the whole European market. As a consequence, some European airports might see a substantial decline in arriving, departing and transfer traffic, thus in revenues and financial solvency, as well as in their connection to other inter-continental and intra-European destinations. At the same time, other airports might realize a significant increase in traffic as they will be sought after by the profit maximizing airlines as their major gateway hubs. Which will be the losing airports and which will be the winning ones? Can airports anticipate the actions of airlines in deregulated markets and utilize policies which will improve their relative position? If so, what should be these anticipatory policies? These questions become the more urgent, since an increasing number of major European airports will be privatized in the near future. Although increasing airport congestion in Europe will also be reflected in a growing demand pressure for airport slots, this is not a guarantee for a stable transport volume growth of individual airports. The more volatile the market is, the more vulnerable privatized airports become. Therefore, the main issue of this study is the analysis of the opportunities of major European airports to become a central hub as a result of the network choices made by the new European airlines in a completely liberalized market. In a previous study (Berechman and de Wit, 1996), we already explored the potential of Amsterdam Airport Schiphol of becoming the major West-European hub, once European aviation markets are deregulated. A major hindrance of that study was the use of a single hub-and-spoke network. For example that model could not analyze the viability of different combinations of European hubs within a multiple hub network of alternative airline alliances. In this study, we have formulated the model of a multi-hub network where two West-European airports are used for inter-continental and intra-European travel to enable a more realistic analysis of hub choice. Like the previous one also this multi-hub model is primarily used to assess the potential ability of Amsterdam Airport Schiphol for becoming a major West-European hub. Thus, in particular, the policy tests focus on this airport in a double hub network.

Berechman, Joseph↗

NASA NTRS1999

Resonant Transmissive Modulator Construction for Use in Beam Steering Array

An agile electronically configurable optical beam steering device for use in directional free-space transmission of optical signals is described. The device design employs an array of tunable resonant transmissive modulators constructed from multi-layered stacks of electro-optically active dielectric materials. Each modulator may be individually configured to transmit an optical signal with a specific amount of phase and group velocity modulation. The resulting diffractive interactions between modulator outputs provide a method for directional optical signal transmission. Of particular focus within this presentation is the optimized design of the individual modulator. It is widely known that periodic multi-layered stacks of dielectric materials exhibit reflective and transmissive resonances as described by the stack's transmission function. If an electric field is applied normal to the stack's layers, phenomena such as the electro-optic effect and/or the quantum-confined Stark effect may be employed to change the effective refractive index within the stack. This refractive index change alters the stack's transmission function and modulates the phase and group velocity of the transmitted optical signal. By varying the parameters that describe the stack's specific construction and operation, such as number of layers within the stack, refractive indexes of stack materials, layer thicknesses, and combinations of periodic verses non-periodic layer repetitions, custom transmittance functions may be generated. The transmitted optical signal carrier frequency and bandwidth must be carefully selected to maximize transmission and to minimize absorption. A computational optimization of the variables describing the stack's construction strives to maximize the amount of optical signal modulation verses applied voltage. Trade-offs between methods of increasing device performance verses the limitations of fabrication technologies are considered.

Keys, Andrew S.↗

NASA NTRS1999

Factors Controlling Stress Rupture of Fiber-Reinforced Ceramic Composites

The successful application of fiber-reinforced ceramic matrix composites (CMC) depends strongly on maximizing material rupture life over a wide range of temperatures and applied stresses. The objective of this paper is to examine the various intrinsic and extrinsic factors that control the high-temperature stress rupture of CMC for stresses below and above those required for cracking of the 0 C plies (Regions I and II, respectively). Using creep-rupture results for a variety of ceramic fibers and rupture data for CMC reinforced by these fibers, it is shown that in those cases where the matrix carries little structural load, CMC rupture conditions can be predicted very well from the fiber behavior measured under the appropriate test environment. As such, one can then examine the intrinsic characteristics of the fibers in order to develop design guidelines for selecting fibers and fiber microstructures in order to maximize CMC rupture life. For those cases where the fiber interfacial coatings are unstable in the test environment, CMC lives are generally worse than those predicted by fiber behavior alone. For those cases where the matrix can support structural load, CMC life can even be greater provided matrix creep behavior is properly controlled. Thus the achievement of long CMC rupture life requires understanding and optimizing the behavior of all constituents in the proper manner.

DiCarlo, J. A.↗

NASA NTRS2000

The Use of HDTV Format and the Electronic Theater in Presenting Earth Science

In order to maximize the public's awareness of earth science observations, earth science data must be available in multiple media formats. This talk will focus on the use High Definition TV format in presenting earth science data, The Television (HDTV) networks are mandated to completely switch over from the current TV standard (NTSC) to HDTV in the next seven years. Museums are also beginning to use HDTV format in their displays. The Visualization Analysis Laboratory at Goddard Space Flight Center has been experimenting with the use of HDTV to present earth science data. The experimental package we have developed is called the Electronic Theater (e-theater). The e-theater is a mobile presentation system used for displaying and teaching groups about earth science and the delicate interdependence between the various earth systems. The e-theater takes advantage of a double-wide screen to show the audiences high resolution data displays. The unique architecture used in this exhibit allows several data sets to be displayed at one time, demonstrating the connections between different earth systems. The data animations are manipulated in real-time during the presentation and can be paused, moved forward, backward, looped, or zoomed into, to maximize the flexibility of the presentation. Because HDTV format is used within the e-theater, the materials generated for the e-theater are made available to the news media and museums.

Summey, Barbara↗

NASA NTRS1999

Cooperation and Coordination Between Fuzzy Reinforcement Learning Agents in Continuous State Partially Observable Markov Decision Processes

Successful operations of future multi-agent intelligent systems require efficient cooperation schemes between agents sharing learning experiences. We consider a pseudo-realistic world in which one or more opportunities appear and disappear in random locations. Agents use fuzzy reinforcement learning to learn which opportunities are most worthy of pursuing based on their promise rewards, expected lifetimes, path lengths and expected path costs. We show that this world is partially observable because the history of an agent influences the distribution of its future states. We consider a cooperation mechanism in which agents share experience by using and-updating one joint behavior policy. We also implement a coordination mechanism for allocating opportunities to different agents in the same world. Our results demonstrate that K cooperative agents each learning in a separate world over N time steps outperform K independent agents each learning in a separate world over K*N time steps, with this result becoming more pronounced as the degree of partial observability in the environment increases. We also show that cooperation between agents learning in the same world decreases performance with respect to independent agents. Since cooperation reduces diversity between agents, we conclude that diversity is a key parameter in the trade off between maximizing utility from cooperation when diversity is low and maximizing utility from competitive coordination when diversity is high.

Berenji, Hamid R.↗