Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Event horizons”

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.

At least 181 records · Page 10

Adaptive cold-load pickup considerations in 2-stage microgrid unit commitment for enhancing microgrid resilience

In an extended main grid outage spanning multiple days, load shedding serves as a critical mechanism for islanded microgrids to maintain essential power and energy reserves that are indispensable for fulfilling reliability and resiliency mandates. However, using load shedding for such purposes leads to increasing occurrence of cold load pickup (CLPU) events. Here, this study presents an innovative adaptive CLPU model that introduces a method for determining and incorporating parameters related to CLPU power and energy requirements into a two-stage microgrid unit commitment (MGUC) algorithm. In contrast to the traditional fixed-CLPU-curve approach, this model calculates CLPU duration, power, and energy demands by considering outage durations and ambient temperature variations within the MGUC process. By integrating the adaptive CLPU model into the MGUC problem formulation, it allows for the optimal allocation of energy resources throughout the entire scheduling horizon to fulfill the CLPU requirements when scheduling multiple CLPU events. The performance of the enhanced MGUC algorithm considering CLPU needs is assessed using actual load and photovoltaic (PV) data. Simulation results demonstrate significant improvements in dispatch optimality evaluated by the amount of load served, customer comfort, energy storage operation, and adherence to energy schedules. These enhancements collectively contribute to reliable and resilient microgrid operation.

24 POWER TRANSMISSION AND DISTRIBUTION↗

High resolution chronology of late Cretaceous-early Tertiary events determined from 21,000 yr orbital-climatic cycles in marine sediments

A number of South Atlantic sites cored by the Deep Sea Drilling Project (DSDP) recovered late Cretaceous and early Tertiary sediments with alternating light-dark, high-low carbonate content. The sedimentary oscillations were turned into time series by digitizing color photographs of core segments at a resolution of about 5 points/cm. Spectral analysis of these records indicates prominent periodicity at 25 to 35 cm in the Cretaceous intervals, and about 15 cm in the early Tertiary sediments. The absolute period of the cycles that is determined from paleomagnetic calibration at two sites is 20,000 to 25,000 yr, and almost certainly corresponds to the period of the earth's precessional cycle. These sequences therefore contain an internal chronometer to measure events across the K/T extinction boundary at this scale of resolution. The orbital metronome was used to address several related questions: the position of the K/T boundary within magnetic chron 29R, the fluxes of biogenic and detrital material to the deep sea immediately before and after the K/T event, the duration of the Sr anomaly, and the level of background climatic variability in the latest Cretaceous time. The carbonate/color cycles that were analyzed contain primary records of ocean carbonate productivity and chemistry, as evidenced by bioturbational mixing of adjacent beds and the weak lithification of the rhythmic sequences. It was concluded that sedimentary sequences that contain orbital cyclicity are capable of providing resolution of dramatic events in earth history with much greater precision than obtainable through radiometric methods. The data show no evidence for a gradual climatic deterioration prior to the K/T extinction event, and argue for a geologically rapid revolution at this horizon.

Herbert, Timothy D.↗

Lightning observations from space shuttle

The experimental program of the Earth Sciences and Applications Division at NASA/MSFC includes development of the Lightning Imaging Sensor (LIS) for the NOAA Earth Observing System (EOS) Polar Platform. The research plan is to use existing lightning information to generate simulated data for the LIS experiment. Navigation algorithms were used to transform pixel locations to latitude and longitude values. The simulated data would then be used to test and develop algorithms for the analysis of LIS data. Individual frames of video imagery obtained from Space Shuttle Missions provide the raw data for the simulation. Individual video frames were digitized to get the pixel locations of lightning flashes. The pixel locations will be used to locate the geographical position of the event. Because of a lack of detailed knowledge of camera orientation with respect to the Space Shuttle, video scenes that contain identifiable city lights were chosen for analysis. A method for locating the payload bay camera axis was developed and tested. Two measurements are needed: the pixel location of the apparent horizon and a timed siting of a known location passing the principal line of the image. Individual video frames were navigated and lightning illuminated clouds were located on the map. Satisfactory agreement in location was achieved for cities and LLP lightning locations. Ground truth measurements were compared to satellite observations. A vertical lightning event was identified on the horizon. Very low frequency (VLF) transmission on this particular occassion shows a strong response to negative cloud to cloud flashes.

Boeck, William L.↗

Resilience Metrics and Framework for Distributed Wind Presentation

This presentation communicates information about the MIRACL project Resilience Metrics report and Resilience Framework report. It was created for the 2021 MIRACL advisory board meeting. We propose a three-tiered approach for the resilience framework. At the top level, we consider the time horizons on which resilience will be evaluated and executed. At the middle level, we consider the core functions of resilience, which span across the time horizons. At the lower level, we consider the process steps that correspond to implementing practices for resilience in each of the core functions. The framework considers three time horizons in order to enable organizations to assess and improve their system’s resilience throughout its lifecycle. We call these time horizons the planning, operational, and future stages. The planning stage uses organizational needs and current system evaluation to prepare for potential hazards. The operational stage seeks to execute responses to hazards as prudently and efficiently as possible to maintain system resilience. The future stage seeks to improve on current system resilience and feeds back into the planning stage to promote continuous improvement. While all three time horizons are important when considering a specific topic, the planning and evaluation phase (i.e., what is done in advance of the event) is critical in defining a system’s resilience characteristics and in outlining how a system responds to an event. This framework intentionally emphasizes the planning stage to highlight the overarching emphasis of this effort, not to imply that the other two time-related horizons (i.e., operational and change the future) are less important. The core functions in the framework are identify, prepare, detect, adapt, and recover. These five functions stem from a rigorous analysis of definitions used across the industry, and they represent the core capabilities that an organization must have to enable lifecycle resilience. Within each core function, process steps are described that help walk an organization through the information gathering, evaluation, decision-making, and implementation processes they will need to ensure their resilience goals are maintained throughout the system and the system lifecycle. Also highlighted in the figure is the concept that a resilience framework should be cyclical in nature. Because a system’s resilience is based on finite resources and time, it must continually evolve through this framework’s risk management and capital investment steps at an appropriate level of scope and pace.

17 WIND ENERGY↗

Data for Rod et al., "Alternating salt and freshwater floods of coastal soils impact soil structure, hydraulic properties, and oxygen dynamics"

This dataset includes laboratory experiment data on soil structure, hydraulic properties, and oxygen dynamics associated with Rod et al. 2026 https://doi.org/10.1002/vzj2.70073. There are six data files from a lab-based flood simulation of either freshwater (FW) or alternating brackish saltwater (SW) and FW using soil cores from a coastal forest at the Smithsonian Environmental Research Center. For soil information please see the Location section of the metadata. Files include: CO2, surface chemistry, water retention, dissolved oxygen, and soil specific surface area. Each file is in CSV format and can be opened/read with any plain text tabular file reader (Microsoft Excel, R, etc.). Purpose of Experiment: To investigate how hydrologic intensification affects soil structure and oxygen dynamics, we conducted a series of laboratory-based flood simulations. After three SW-FW floods (6 floods total) there were significant changes in pore size distribution, significant redistribution of colloids, and the A-horizon became sodic. We concluded that a small number of SW flooding events can induce a measurable change in soil physical properties that directly impacts the biogeochemical dynamics.

54 ENVIRONMENTAL SCIENCES↗

Survival analysis of localized prostate cancer with deep learning

In recent years, data-driven, deep-learning-based models have shown great promise in medical risk prediction. By utilizing the large-scale Electronic Health Record data found in the U.S. Department of Veterans Affairs, the largest integrated healthcare system in the United States, we have developed an automated, personalized risk prediction model to support the clinical decision-making process for localized prostate cancer patients. This method combines the representative power of deep learning and the analytical interpretability of parametric regression models and can implement both time-dependent and static input data. To collect a comprehensive evaluation of model performances, we calculate time-dependent C-statistics C td over 2-, 5-, and 10-year time horizons using either a composite outcome or prostate cancer mortality as the target event. The composite outcome combines the Prostate-Specific Antigen (PSA) test, metastasis, and prostate cancer mortality. Our longitudinal model Recurrent Deep Survival Machine (RDSM) achieved C td 0.85 (0.83), 0.80 (0.83), and 0.76 (0.81), while the cross-sectional model Deep Survival Machine (DSM) attained C td 0.85 (0.82), 0.80 (0.82), and 0.76 (0.79) for the 2-, 5-, and 10-year composite (mortality) outcomes, respectively. In addition to estimating the survival probability, our method can quantify the uncertainty associated with the prediction. The uncertainty scores show a consistent correlation with the prediction accuracy. We find PSA and prostate cancer stage information are the most important indicators in risk prediction. Our work demonstrates the utility of the data-driven machine learning model in prostate cancer risk prediction, which can play a critical role in the clinical decision system.

60 APPLIED LIFE SCIENCES↗

Prospective Safety Analysis and the Complex Aviation System

Fatal accident rates in commercial passenger aviation are at historic lows yet have plateaued and are not showing evidence of further safety advances. Modern aircraft accidents reflect both historic causal factors and new unexpected "Black Swan" events. The ever-increasing complexity of the aviation system, along with its associated technology and organizational relationships, provides fertile ground for fresh problems. It is important to take a proactive approach to aviation safety by working to identify novel causation mechanisms for future aviation accidents before they happen. Progress has been made in using of historic data to identify the telltale signals preceding aviation accidents and incidents, using the large repositories of discrete and continuous data on aircraft and air traffic control performance and information reported by front-line personnel. Nevertheless, the aviation community is increasingly embracing predictive approaches to aviation safety. The "prospective workshop" early assessment tool described in this paper represents an approach toward this prospective mindset-one that attempts to identify the future vectors of aviation and asks the question: "What haven't we considered in our current safety assessments?" New causation mechanisms threatening aviation safety will arise in the future because new (or revised) systems and procedures will have to be used under future contextual conditions that have not been properly anticipated. Many simulation models exist for demonstrating the safety cases of new operational concepts and technologies. However the results from such models can only be as valid as the accuracy and completeness of assumptions made about the future context in which the new operational concepts and/or technologies will be immersed. Of course that future has not happened yet. What is needed is a reasonably high-confidence description of the future operational context, capturing critical contextual characteristics that modulate both the likelihood of occurrence of hazards, and the likelihood that those hazards will lead to negative safety events. Heuristics extracted from scenarios, questionnaires, and observed trends from scanning the aviation horizon may be helpful in capturing those future changes in a way conducive to safety assessment. What is also needed is a checklist of potential sources of emerging risk that arise from organizational features that are frequently overlooked. The ultimate goal is to develop a pragmatic, workable method for using descriptions of the future aviation context, to generate valid predictions of safety risks.

prospection↗

Propagating synthetic populations with dynamic Bayesian networks: a framework for long-horizon demographic forecasting

This study presents a dynamic demographic microsimulator using dynamic Bayesian networks to forecast long–term changes in household and individual life events. Leveraging longitudinal Panel Study of Income Dynamics (PSID) data, two networks for individuals and households were modeled to simulate transitions in employment, income, education, marriage, childbirth, leaving the parental home, home ownership, mortality, and household formation or dissolution. Across 1,000 simulation runs spanning 24 years, household–level outcomes remain highly accurate and individual–level predictions reasonable. Although accuracy naturally declines with projection horizon, performance remains promising at both levels. This study addresses a key limitation of existing population synthesis models, which typically generate only a single static snapshot of the population. In conclusion, by introducing a framework that propagates cross-sectional outputs into the future, the microsimulator enables the tracking of demographic evolution over time, enhances realism in population-based simulations, and supplies credible inputs to agent-based travel demand models.

Demographic modeling↗

Analyzing Launch Operations using the Spaceport Throughput Analysis Resource (STAR)

NASA’s Kennedy Space Center (KSC) and Cape CanaveralSpace Force Station (CCSFS) are the world’s preeminent multi-user spaceport, providing facilities and launch capabilities to the agency, NASA’s commercial partners, and other government agencies. KSC works to ensure an environment in which NASA’s programs and other users can safely and effectively carry out their operations. Approximately 100 launches are expected to take place in 2024. KSC needs to meet the needs of customers who request launches based on their own schedules, but who may not be aware of KSC-wide resource limitations and external constraints on launch operations. KSC-wide resources include telecommunications, range tracking, commodities such as Helium and Nitrogen, and special equipment needed to support launches. External constraints include seasonal launch operations. Since space vehicles are complex and launch operations are uncertain, unexpected events can also cause delays in operations. The Spaceport Throughput AnalysisResource (STAR) assesses whether a specific set of pro-posed launches and associated activities can be performed given the resources KSC currently has available, and external constraints imposed on KSC operations. STAR integrates short-horizon scheduling and simulation of launches using a monte-carlo approach driven by configurable probabilities of different classes of event outcomes, including de-lays and worst-case use of resources. The resulting analyses inform stakeholders of key constraints preventing customers from being able to perform their missions as desired, and give insight into how to improve spaceport throughput. The rest of the paper is organized as follows. In 2 we de-scribe the problem STAR is intended to solve. In Section3 we formally describe the scheduling problem ingredients. In Section 4 we describe how manifests are simulated. InSection 5 we describe the constraints problem that is solved when rescheduling. In Section 6 we describe what information is recorded during manifest simulation in order to evaluate delays. In Section 7 we describe the specific problem solved for KSC. In Section 8 we describe the challenges of knowledge engineering for STAR.

Planning Scheduling Analysis↗

Paleo-Environment and C-14 Dating: The Key to the Depositional Age of the Tha Chang and Related Sand Pits, Northeastern Thailand

Tha Chang sand pits, Nakhon Ratchasima Province and many other sand pits in the area adjacent to the Mun River are characterized by their fluviatile environment in association with mass wasting deposits, along the paleo-river channel and the flood plain of the Mun River. Sediments of these deposits are characterized by clasts of various rock types especially the resistant ones with frequent big tree trunks, logs and wood fragments in different sizes and various stages of transformation from moldering stage to lignification and petrification. Widespread pyritization of the lower horizon suggests strongly reducing environment during burial. The Tha Chang deposits have been received much attention from geoscientists especially paleontologist communities, as they contain fragments of some distinct vertebrate species such as Stegadon sp., hominoid primate, rhinoceros Aceratherium and others. Based on the associated mammal fauna and hominoid fossils, the late Miocene ( 9 - 6 Ma) was given for the time of deposition of this sand and gravel unit. Some other reports believed that sediments and materials of these sand and gravel quarries (pits) were deposited by high-energy flood pulses contemporaneous with the tektites forming event during mid-Pleistocene at c. 0.8 Ma. Interpretation from Palynostratigraphical study suggested that the lower horizon of Tha Chang sand pit was deposited during Pliocene/Pleistocene period and the upper horizons are Pleistoncene/Holocene. It is crystal clear that all the fluviatile sediments including tektites and almost all fossil fragments being deposited in these sand pits were, likely a multiple times reworked materials. Only some old bamboo trees, some old crowling trees and fossils grasses observed on the old river bank are considered in situ. C-14 dating of 5 old wood specimens from Tha Chang Sand Pits, 15 old wood specimens from Chumpuang Sand Pits and one sample of old pottery from a Chumpuang Sand Pit were carried out in the NSF- Arizona AMS Laboratory. Although, there is no sharp boundary between the unconsolidated sedimentary horizons in the pits, C-14 ages obtained from the Tha Chang vary from 34,340 BP at the middle horizon (approx 10 m below ground zero) to >49,900 BP at the lower horizon with unknown basal formation (highly pyritized zone approx 20 - 25 m below ground zero). The ages for the Chumpuang vary from 41,700 BP, >45,900 BP and >49,900 BP from the upper most to the lower most of a broad horizon (approx 8 m to approx 12 m below ground zero). The C-14 age of the pottery collected from layer approximately 5 m below ground zero is 2,514 BP. The nature of fluviatile together with occasional mass wasting characteristics of all sand pits studies suggest the relatively faster depositional rate of the lower horizon which involved more flooding and mass wasting deposits than those of the upper horizons. The apparent of some mixing of the wood ages may indicate reworking and lag deposits nature of the area. The depositional rate of the upper most sand and soil horizon (5 m thick) is approximately 1 m per 500 years which mean both erosion and deposition had played a significant role during that time period. In term of the true age of the formation, we argue that since most of the materials deposited are reworked materials, all ages obtained from fossil fragments could not be the age of sand and gravel formation. Furthermore, the maximum age of all the tektite bearing horizons cannot be older than 0.8 Ma. The oldest C-14 age of 49,900 BP is interpreted as the minimum age of the Tha Chang and related sand pits formation when geomorphology of the area was a lot more hilly and much higher gradient than that of the present day.

Putthapiban, P.↗

Demonstrating Acquisition of Real-Time Thermal Data over Fires Utilizing UAVs

NASA-Ames Research Center, in collaboration with General Atomics Aeronautical Systems, Inc. has been developing real-time data acquisition and information delivery systems employing uninhabited aerial vehicle (UAV) technology for disaster mitigation and assessment demonstrations. Working in conjunction with the US Forest Service, a disaster community agency responsible for wildfire management and mitigation, we developed a large-scale wildfire demonstration called the First Response Experiment (FIRE). During that experiment in late summer 2001, the participants demonstrated the melding of innovative technologies such as UAV platforms, real-time data processing, and data telemetry for quick analysis of a disaster event. The General Atomics ALTUS UAV, the Airborne Infrared Disaster Assessment System (AIRDAS) and Over-The-Horizon (OTH) satellite data telemetry equipment were employed over a controlled burn to test the feasibility of a disaster monitoring and mitigation platform for hazardous duty. The ALTUS UAV was employed to demonstrate the long duration, altitude, and payload capability of unmanned platforms for acquiring disaster related data. The ALTUS has an operational altitude to 45,000 feet (13,700 in), with a flight duration of twenty-four hours and a payload capacity of over 300 lbs. (148.5 kg). This allows the platform to operate under the conditions that would be necessary for monitoring and mitigating disaster events throughout the Unites States. The four channel AIRDAS data (calibrated thermal infrared digital imagery of the fire event) was sent from the ALTUS UAV via a satellite communications system (NERA transponder and INMARSAT satellite) to a data archive server and an image processing work station at NASA-Ames Research Center, 400 miles away.

Wegener, Steven S.↗

X-ray Transients in the Advanced LIGO/Virgo Horizon

Advanced LIGO and Advanced Virgo will be all-sky monitors for merging compact objects within a few hundred megaparsecs. Finding the electromagnetic counterparts to these events will require an understanding of the transient sky at low redshift (z < 0.1). We performed a systematic search for extragalactic, low redshift, transient events in the XMM-Newton Slew Survey. In a flux limited sample, we found that highly variable objects comprised 10% of the sample, and that of these, 10% were spatially coincident with cataloged optical galaxies. This led to 4 × 10(exp −4) transients per square degree above a flux threshold of 3×10(exp −12) erg/sq cm/s (0.2-2 keV) which might be confused with LIGO/Virgo counterparts. This represents the first extragalactic measurement of the soft X-ray transient rate within the Advanced LIGO/Virgo horizon. Our search revealed six objects that were spatially coincident with previously cataloged galaxies, lacked evidence for optical active galactic nuclei, displayed high luminosities approx. 10(exp 43) erg/s, and varied in flux by more than a factor of 10 when compared with the ROSAT All-Sky Survey. At least four of these displayed properties consistent with previously observed tidal disruption events.

gravitational waves↗

Long-Range Transhorizon Lunar Surface Radio Wave Propagation in the Presence of a Regolith and a Sparse Exospheric Plasma

Long-range, over-the-horizon (transhorizon) radio wave propagation is considered for the case of the Moon. In the event that relay satellites are not available or otherwise unwarranted for use, transhorizon communication provides for a contingency or backup option for non line-of-sight lunar surface exploration scenarios. Two potential low-frequency propagation mechanisms characteristic of the lunar landscape are the lunar regolith and the photoelectron induced plasma exosphere enveloping the Moon. Although it was hoped that the regolith would provide for a spherical waveguide which could support a trapped surface wave phenomena, it is found that, in most cases, the regolith is deleterious to long range radio wave propagation. However, the presence of the plasma of the lunar exosphere supports wave propagation and, in fact, surpasses the attenuation of the regolith. Given the models of the regolith and exosphere adopted here, it is recommended that a frequency of 1 MHz be considered for low rate data transmission along the lunar surface. It is also recommended that further research be done to capture the descriptive physics of the regolith and the exospheric plasma so that a more complete model can be obtained. This comprehensive theoretical study is based entirely on first principles and the mathematical techniques needed are developed as required; it is self-contained and should not require the use of outside resources for its understanding.

Manning, Robert M.↗

Cryovolcanic flooding in Viking Terra on Pluto

A prominent fossa trough (Uncama Fossa) and adjacent 28-km diameter impact crater (Hardie) in Pluto's Viking Terra, as seen in the high-resolution images from the New Horizons spacecraft, show morphological evidence of in-filling with a material of uniform texture and red-brown color. A linear fissure parallel to the trough may be the source of a fountaining event yielding a cryoclastic deposit having the same composition and color properties as is found in the trough and crater. Spectral maps of this region with the New Horizons LEISA instrument reveal the spectral signature of H2O ice in these structures and in distributed patches in the adjacent terrain in Viking Terra. A detailed statistical analysis of the spectral maps shows that the colored H2O ice filling material also carries the 2.2-μm signature of an ammoniated component that may be an ammonia hydrate (NH3·nH2O) or an ammoniated salt. This paper advances the view that the crater and fossa trough have been flooded by a cryolava debouched from Pluto's interior along fault lines in the trough and in the floor of the impact crater. The now frozen cryolava consisted of liquid H2O infused with the red-brown pigment presumed to be a tholin, and one or more ammoniated compounds. Although the abundances of the pigment and ammoniated compounds entrained in, or possibly covering, the H2O ice are unknown, the strong spectral bands of the H2O ice are clearly visible. In consideration of the factors in Pluto's space environment that are known to destroy ammonia and ammonia-water mixtures, the age of the exposure is of order ≤109 years. Ammoniated salts may be more robust, and laboratory investigations of these compounds are needed.

Pluto↗

Soil moisture variation patterns observed in Hand County, South Dakota

Soil moisture data were taken during 1976 (April, June, October), 1977 (April, May, June), and 1978 (May, June, July) Hand County, South Dakota as part of the ground truth used in NASA's aircraft experiments to study the use of microwave radiometers for the remote sensing of soil moisture. The spatial variability observed on the ground during each of the sampling events was studied. The data reported are the mean gravimetric soil moisture contained in three surface horizon depths: 0 to 2.5, 0 to 5 and 0 to 10 cm. The overall moisture levels ranged from extremely dry conditions in June 1976 to very wet in May 1978, with a relatively even distribution of values within that range. It is indicated that well drained sites have to be partitioned from imperfectly drained areas when attempting to characterize the general moisture profile throughout an area of varying soil and cover type conditions. It is also found that the variability in moisture content is greatest in the 0 to 2.5 cm measurements and decreases as the measurements are integrated over a greater depth. It is also determined that the sampling intensity of 10 measurements per km is adequate to estimate the mean moisture with an uncertainty of + or - 3 percent under average moisture conditions in areas of moderate to good drainage.

Jones, E. B.↗

The Upgrade of Horizon-T Detector

The Horizon-T experiment is located at the elevation of 3346 m above sea level near the city of Almaty, Republic of Kazakhstan. A thorough comparison of the spatial and temporal characteristics of charged components of Extended Air Showers (EAS) with delayed particles has been conducted between the simulated EAS using CORSIKA simulation package and the selection from the experimental data set of events with two pulses recorded by a detector at ~600 m distance from axis [1]. This comparison has shown that events with delayed particles cannot be described within existing simulation models. The significance of these results prompted the upgrade of the Horizon-T experiment. New detector points have been added at the ~600m to enhance data at that distance. Fast glass-based detector has been added to the scintillator-based detector at center point for accurate measurements of the pulse widths.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Low-Latitude Auroras: The Magnetic Storm of 14-15 May 1921

We review solar geophysical data relating to the great magnetic storm of 14-15 May 1921, with emphasis on observations of the low-latitude visual aurora. From the reports we have gathered for this event the lowest geomagnetic latitude of definite overhead aurora (coronal form) was 40 deg and the lowest geomagnetic latitude from which auroras were observed on the poleward horizon in the northern hemisphere was 30 deg. For comparison, corresponding overhead/low-latitude values of 48 deg/32 deg and 41 deg/20 deg were reported for the great auroras on 28-29 August and 1-2 September 1859, respectively. However for the 1921 event, there is a report of aurora from Apia, Samoa, in the southern hemisphere, within 13 deg of the geomagnetic equator. This report by professional observers appears to be credible, based on the aurora description and timing, but is puzzling because of the discrepancy with the lowest latitude of observation in the northern hemisphere and the great implied aurora height (approximately 2000 km, assuming overhead aurora at Auckland, New Zealand). We discuss various possibilities that might account for this observation.

Silverman, S. M.↗

Reinforcement Learning in a Nonstationary Environment: The El Farol Problem

This paper examines the performance of simple learning rules in a complex adaptive system based on a coordination problem modeled on the El Farol problem. The key features of the El Farol problem are that it typically involves a medium number of agents and that agents' pay-off functions have a discontinuous response to increased congestion. First we consider a single adaptive agent facing a stationary environment. We demonstrate that the simple learning rules proposed by Roth and Er'ev can be extremely sensitive to small changes in the initial conditions and that events early in a simulation can affect the performance of the rule over a relatively long time horizon. In contrast, a reinforcement learning rule based on standard practice in the computer science literature converges rapidly and robustly. The situation is reversed when multiple adaptive agents interact: the RE algorithms often converge rapidly to a stable average aggregate attendance despite the slow and erratic behavior of individual learners, while the CS based learners frequently over-attend in the early and intermediate terms. The symmetric mixed strategy equilibria is unstable: all three learning rules ultimately tend towards pure strategies or stabilize in the medium term at non-equilibrium probabilities of attendance. The brittleness of the algorithms in different contexts emphasize the importance of thorough and thoughtful examination of simulation-based results.

Bell, Ann Maria↗