SEARCH · Search NASA

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 559 records · Page 31

All sources NASA NTRS DOE OSTI

NASA NTRS1977

Optimum antireflection coating for Antireflection-coated Metal-Oxide-Semiconductor /AMOS/ solar cells

Consideration is given to the design of a single-layer optimum antireflection coating for AMOS (antireflection-coated metal-oxide-semiconductor) solar cells to match the entire sunlight spectrum. The energy conversion efficiency is maximized by maximizing the open-circuit voltage and the short-circuit current. The former is maximized by oxidation techniques and the latter is maximized by the light-coupling into the solar cell. With reference to the effective index of refraction as obtained by ellipsometry, examples of optimum antireflection coatings for 60-A Au-GaAs solar cells are presented.

Yeh, Y. C. M.↗

NASA NTRS1990

Additive evaluation criteria for aircraft noise

The occurrence of unexpected aircraft noise events will frequently evoke intense complaints about annoyance over such events. It is recognized that the relationship between the volume of complaints and the corresponding maximum noise levels does in fact depend on the circumstances of the complainants and the time of year. The frequency of occurrence of the respective noise events is also a factor. The possible practical value of the addition of the maximal noise level, to the well-known cumulative noise descriptors L(sub eq), L(sub dn), etc. may be considered. One might start by considering the difference between the L(sub eq) and the average maximal noise level of the twenty loudest single noise events on an average day. A somewhat less sharply focused consideration of the maximum noise levels was adopted. In that standard the scope of noise-mitigation measures is defined generally with reference to L(sub eq). If the average maximal noise level L(sub max) of the entire aircraft fleet mix exceeds the L(sub eq) by more than 20 dB(A) and if more than 20 daily aircraft noise events exceed the L(sub eq) than the difference L(sub max)-20 becomes the key criterion for noise-mitigation measures. A statistical correlation of a large number of data from aircraft-noise-monitoring sensors located both directly underneath and laterally disposed to an aircraft flight-path has supplied a basis for the determination of the distribution of the maximal noise levels about the average value, L(sub max), of each type of aircraft. This distribution is given for the takeoff climb and for the landing approach.

Meyer, Thomas J.↗

NASA NTRS1986

Exercise Responses after Inactivity

The exercise response after bed rest inactivity is a reduction in the physical work capacity and is manifested by significant decreases in oxygen uptake. The magnitude of decrease in maximal oxygen intake V(dot)O2max is related to the duration of confinement and the pre-bed-rest level of aerobic fitness; these relationships are relatively independent of age and gender. The reduced exercise performance and V(dot)O2max following bed rest are associated with various physiological adaptations including reductions in blood volume, submaximal and maximal stroke volume, maximal cardiac output, sceletal muscle tone and strength, and aerobic enzyme capacities, as well as increases in venous compliance and submaximal and maximal heart rate. This reduction in physiological capacity can be partially restored by specific countermeasures that provide regular muscular activity or orhtostatic stress or both during the bed rest exposure. The understanding of these physiological and physical responses to exercise following bed rest inactivity has important implications for the solution to safety and health problems that arise in clinical medicine, aerospace medicine, sedentary living, and aging.

Convertino, Victor A.↗

NASA NTRS2000

Inflight Exercise Regimen for the 2-Hour Prebreathe Protocol

A 10 min aerobic prebreathe exercise up to 75% V-O2(sub max) on a dual-cycle ergometer, included in the 2-hour prebreathe protocol, has been shown to dramatically reduce the incidence of decompression sickness (DCS) at altitude. In-flight only leg ergometry will be available. A balanced exercise was developed using surgical tubing with the ergometer on-orbit. We hypothesize that a 75% V02max workload, individually prescribed, would be achieved using a target heart rate to regulate the intensity of the arm exercise. VO2, heart rate (HR) / ECG, V-CO2 /V-O2, V(sub E), and V(sub T), and rate of perceived exertion (Borg scale) were measured in eleven healthy subjects who passed a US Air Force Class III Physical examination. A V-O2 peak test was performed to assess the sub-maximal exercise prescription. Two series of sub-maximal tests were performed: (1) leg ergometer/hand ergometer and (2) leg ergometer/surgical tubes. We found no significant differences (P > 0.05) in comparing the means for V-O2 and HR between the predicted and measured values during the final 4 minute-stage at "75% V-O2 workload" or between the two types of sub-maximal tests. The prescribed prebreathe sub-maximal exercise performed with flight certified surgical tubes was achieved using the target HR.

Foster, Philip P.↗

NASA NTRS2002

Heterogeneous responses of human limbs to infused adrenergic agonists: a gravitational effect?

Unlike quadrupeds, the legs of humans are regularly exposed to elevated pressures relative to the arms. We hypothesized that this "dependent hypertension" would be associated with altered adrenergic responsiveness. Isoproterenol (0.75-24 ng x 100 ml limb volume-1 x min-1) and phenylephrine (0.025-0.8 microg x 100 ml limb volume-1 x min-1) were infused incrementally in the brachial and femoral arteries of 12 normal volunteers; changes in limb blood flow were quantified by using strain-gauge plethysmography. Compared with the forearm, baseline calf vascular resistance was greater (38.8 +/- 2.5 vs. 26.9 +/- 2.0 mmHg x 100 ml x min x ml-1; P < 0.001) and maximal conductance was lower (46.1 +/- 11.9 vs. 59.4 +/- 13.4 ml x ml-1 x min-1 x mmHg-1; P < 0.03). Vascular conductance did not differ between the two limbs during isoproterenol infusions, whereas decreases in vascular conductance were greater in the calf than the forearm during phenylephrine infusions (P < 0.001). With responses normalized to maximal conductance, the half-maximal response for phenylephrine was significantly less for the calf than the forearm (P < 0.001), whereas the half-maximal response for isoproterenol did not differ between limbs. We conclude that alpha1- but not beta-adrenergic-receptor responsiveness in human limbs is nonuniform. The relatively greater response to alpha1-adrenergic-receptor stimulation in the calf may represent an adaptive mechanism that limits blood pooling and capillary filtration in the legs during standing.

Clinical Trial↗

NASA NTRS1997

Response properties of pigeon otolith afferents to linear acceleration

In the present study, the sensitivity to sinusoidal linear accelerations in the plane of the utricular macula was tested in afferents. The head orientation relative to the translation axis was varied in order to determine the head position that elicited the maximal and minimal responses for each afferent. The response gain and phase values obtained to 0.5-Hz and 2-Hz linear acceleration stimuli were then plotted as a function of head orientation and a modified cosine function was fit to the data. From the best-fit cosine function, the predicted head orientations that would produce the maximal and minimal response gains were estimated. The estimated maximum response gains to linear acceleration in the utricular plane for the afferents varied between 75 and 1420 spikes s-1 g-1. The mean maximal gains for all afferents to 0.5-Hz and 2-Hz sinusoidal linear acceleration stimuli were 282 and 367 spikes s-1 g-1, respectively. The minimal response gains were essentially zero for most units. The response phases always led linear acceleration and remained constant for each afferent, regardless of head orientation. These response characteristics indicate that otolith afferents are cosine tuned and behave as one-dimensional linear accelerometers. The directions of maximal sensitivity to linear acceleration for the afferents varied throughout the plane of the utricle; however, most vectors were directed out of the opposite ear near the interaural axis. The response dynamics of the afferents were tested using stimulus frequencies ranging between 0.25 Hz and 10 Hz (0.1 g peak acceleration). Across stimulus frequencies, most afferents had increasing gains and constant phase values. These dynamic properties for individual afferents were fit with a simple transfer function that included three parameters: a mechanical time constant, a gain constant, and a fractional order distributed adaptation operator.

Non-NASA Center↗

NASA NTRS2004

Environmental impacts on the developing CNS: CD15, NCAM-L1, and GFAP expression in rat neonates exposed to hypergravity

We have previously reported that the developing rat cerebellum is affected by hypergravity exposure. The effect is observed during a period of both granule and glial cell proliferation and neuronal migration in the cerebellum and coincides with changes in thyroid hormone levels. The present study begins to address the molecular mechanisms involved in the cerebellar response to hypergravity. Specifically, the study focuses on the expression of cerebellar proteins that are known to be directly involved in cell-cell interactions [protein expressing 3-fucosyl-N-acetyl-lactosamine antigen (CD15), neuronal cell adhesion molecule (NCAM-L1)] and those that affect cell-cell interactions indirectly [glial fibrillary acidic protein (GFAP)] in rat neonates exposed to centrifuge-produced hypergravity. Cerebellar mass and protein expression in rat neonates exposed to hypergravity (1.5 G) from gestational day (G) 11 to postnatal day (P) 30 were compared at one of six time points between P6 and P30 against rat neonates developing under normal gravity. Proteins were analyzed by quantitative western blots of cerebellar homogenates prepared from male or female neonates. Cerebellar size was most clearly reduced in male neonates on P6 and in female neonates on P9, with a significant gender difference; differences in cerebellar mass remained significant even when change in total body mass was factored in. Densitometric analysis of western blots revealed both quantitative and temporal changes in the expression of selected cerebellar proteins that coincided with changes in cerebellar mass and were gender-specific. In fact, our data indicated certain significant differences even between male and female control animals. A maximal decrease in expression of CD15 was observed in HG females on P9, coinciding with maximal change in their cerebellar mass. A shift in the time-course of NCAM-L1 expression resulted in a significant increase in NCAM-L1 in HG males on P18, an isolated time at which cerebellar mass does not significantly differ between HG and SC neonates. A maximal decrease in expression of GFAP was observed in HG males on P6, coinciding with maximal change in their cerebellar mass. Altered expression of cerebellar proteins is likely to affect a number of developmental processes and contribute to the structural and functional alterations seen in the CNS developing under altered gravity. Our data suggest that both cerebellar development and its response to gravitational manipulations differ in males and females. c2004 COSPAR. Published by Elsevier Ltd. All rights reserved.

NASA Program Fundamental Space Biology↗

NASA NTRS2006

Steps Toward Optimal Competitive Scheduling

This paper is concerned with the problem of allocating a unit capacity resource to multiple users within a pre-defined time period. The resource is indivisible, so that at most one user can use it at each time instance. However, different users may use it at different times. The users have independent, se@sh preferences for when and for how long they are allocated this resource. Thus, they value different resource access durations differently, and they value different time slots differently. We seek an optimal allocation schedule for this resource. This problem arises in many institutional settings where, e.g., different departments, agencies, or personal, compete for a single resource. We are particularly motivated by the problem of scheduling NASA's Deep Space Satellite Network (DSN) among different users within NASA. Access to DSN is needed for transmitting data from various space missions to Earth. Each mission has different needs for DSN time, depending on satellite and planetary orbits. Typically, the DSN is over-subscribed, in that not all missions will be allocated as much time as they want. This leads to various inefficiencies - missions spend much time and resource lobbying for their time, often exaggerating their needs. NASA, on the other hand, would like to make optimal use of this resource, ensuring that the good for NASA is maximized. This raises the thorny problem of how to measure the utility to NASA of each allocation. In the typical case, it is difficult for the central agency, NASA in our case, to assess the value of each interval to each user - this is really only known to the users who understand their needs. Thus, our problem is more precisely formulated as follows: find an allocation schedule for the resource that maximizes the sum of users preferences, when the preference values are private information of the users. We bypass this problem by making the assumptions that one can assign money to customers. This assumption is reasonable; a committee is usually in charge of deciding the priority of each mission competing for access to the DSN within a time period while scheduling. Instead, we can assume that the committee assigns a budget to each mission.This paper is concerned with the problem of allocating a unit capacity resource to multiple users within a pre-defined time period. The resource is indivisible, so that at most one user can use it at each time instance. However, different users may use it at different times. The users have independent, se@sh preferences for when and for how long they are allocated this resource. Thus, they value different resource access durations differently, and they value different time slots differently. We seek an optimal allocation schedule for this resource. This problem arises in many institutional settings where, e.g., different departments, agencies, or personal, compete for a single resource. We are particularly motivated by the problem of scheduling NASA's Deep Space Satellite Network (DSN) among different users within NASA. Access to DSN is needed for transmitting data from various space missions to Earth. Each mission has different needs for DSN time, depending on satellite and planetary orbits. Typically, the DSN is over-subscribed, in that not all missions will be allocated as much time as they want. This leads to various inefficiencies - missions spend much time and resource lobbying for their time, often exaggerating their needs. NASA, on the other hand, would like to make optimal use of this resource, ensuring that the good for NASA is maximized. This raises the thorny problem of how to measure the utility to NASA of each allocation. In the typical case, it is difficult for the central agency, NASA in our case, to assess the value of each interval to each user - this is really only known to the users who understand their needs. Thus, our problem is more precisely formulated as follows: find an allocation schedule for the resource that maximizes the sum ofsers preferences, when the preference values are private information of the users. We bypass this problem by making the assumptions that one can assign money to customers. This assumption is reasonable; a committee is usually in charge of deciding the priority of each mission competing for access to the DSN within a time period while scheduling. Instead, we can assume that the committee assigns a budget to each mission.

Frank, Jeremy↗

NASA NTRS2008

Optimized Geometry for Superconducting Sensing Coils

An optimized geometry has been proposed for superconducting sensing coils that are used in conjunction with superconducting quantum interference devices (SQUIDs) in magnetic resonance imaging (MRI), magnetoencephalography (MEG), and related applications in which magnetic fields of small dipoles are detected. In designing a coil of this type, as in designing other sensing coils, one seeks to maximize the sensitivity of the detector of which the coil is a part, subject to geometric constraints arising from the proximity of other required equipment. In MRI or MEG, the main benefit of maximizing the sensitivity would be to enable minimization of measurement time. In general, to maximize the sensitivity of a detector based on a sensing coil coupled with a SQUID sensor, it is necessary to maximize the magnetic flux enclosed by the sensing coil while minimizing the self-inductance of this coil. Simply making the coil larger may increase its self-inductance and does not necessarily increase sensitivity because it also effectively increases the distance from the sample that contains the source of the signal that one seeks to detect. Additional constraints on the size and shape of the coil and on the distance from the sample arise from the fact that the sample is at room temperature but the coil and the SQUID sensor must be enclosed within a cryogenic shield to maintain superconductivity.

Eom, Byeong Ho↗

NASA NTRS2010

Practical Loop-Shaping Design of Feedback Control Systems

An improved methodology for designing feedback control systems has been developed based on systematically shaping the loop gain of the system to meet performance requirements such as stability margins, disturbance attenuation, and transient response, while taking into account the actuation system limitations such as actuation rates and range. Loop-shaping for controls design is not new, but past techniques do not directly address how to systematically design the controller to maximize its performance. As a result, classical feedback control systems are designed predominantly using ad hoc control design approaches such as proportional integral derivative (PID), normally satisfied when a workable solution is achieved, without a good understanding of how to maximize the effectiveness of the control design in terms of competing performance requirements, in relation to the limitations of the plant design. The conception of this improved methodology was motivated by challenges in designing control systems of the types needed for supersonic propulsion. But the methodology is generally applicable to any classical control-system design where the transfer function of the plant is known or can be evaluated. In the case of a supersonic aerospace vehicle, a major challenge is to design the system to attenuate anticipated external and internal disturbances, using such actuators as fuel injectors and valves, bypass doors, and ramps, all of which are subject to limitations in actuator response, rates, and ranges. Also, for supersonic vehicles, with long slim type of structures, coupling between the engine and the structural dynamics can produce undesirable effects that could adversely affect vehicle stability and ride quality. In order to design distributed controls that can suppress these potential adverse effects, within the full capabilities of the actuation system, it is important to employ a systematic control design methodology such as this that can maximize the effectiveness of the control design in a methodical and quantifiable way. The emphasis is in generating simple but rather powerful design techniques that will allow even designers with a layman s knowledge in controls to develop effective feedback control designs. Unlike conventional ad hoc methodologies of feedback control design, in this approach actuator rates are incorporated into the design right from the start: The relation between actuator speeds and the desired control bandwidth of the system is established explicitly. The technique developed is demonstrated via design examples in a step-by-step tutorial way. Given the actuation system rates and range limits together with design specifications in terms of stability margins, disturbance rejection, and transient response, the procedure involves designing the feedback loop gain to meet the requirements and maximizing the control system effectiveness, without exceeding the actuation system limits and saturating the controller. Then knowing the plant transfer function, the procedure involves designing the controller so that the controller transfer function together with the plant transfer function equate to the designed loop gain. The technique also shows what the limitations of the controller design are and how to trade competing design requirements such as stability margins and disturbance rejection. Finally, the technique is contrasted against other more familiar control design techniques, like PID control, to show its advantages.

Kopasakis, George↗

NASA NTRS2011

Turboelectric Distributed Propulsion in a Hybrid Wing Body Aircraft

The performance of the N3-X, a 300 passenger hybrid wing body (HWB) aircraft with turboelectric distributed propulsion (TeDP), has been analyzed to see if it can meet the 70% fuel burn reduction goal of the NASA Subsonic Fixed Wing project for N+3 generation aircraft. The TeDP system utilizes superconducting electric generators, motors and transmission lines to allow the power producing and thrust producing portions of the system to be widely separated. It also allows a small number of large turboshaft engines to drive any number of propulsors. On the N3-X these new degrees of freedom were used to (1) place two large turboshaft engines driving generators in freestream conditions to maximize thermal efficiency and (2) to embed a broad continuous array of 15 motor driven propulsors on the upper surface of the aircraft near the trailing edge. That location maximizes the amount of the boundary layer ingested and thus maximizes propulsive efficiency. The Boeing B777-200LR flying 7500 nm (13890 km) with a cruise speed of Mach 0.84 and an 118100 lb payload was selected as the reference aircraft and mission for this study. In order to distinguish between improvements due to technology and aircraft configuration changes from those due to the propulsion configuration changes, an intermediate configuration was included in this study. In this configuration a pylon mounted, ultra high bypass (UHB) geared turbofan engine with identical propulsion technology was integrated into the same hybrid wing body airframe. That aircraft achieved a 52% reduction in mission fuel burn relative to the reference aircraft. The N3-X was able to achieve a reduction of 70% and 72% (depending on the cooling system) relative to the reference aircraft. The additional 18% - 20% reduction in the mission fuel burn can therefore be attributed to the additional degrees of freedom in the propulsion system configuration afforded by the TeDP system that eliminates nacelle and pylon drag, maximizes boundary layer ingestion (BLI) to reduce inlet drag on the propulsion system, and reduces the wake drag of the vehicle.

Felder, James L.↗

NASA NTRS2018

Evaluation of the Danish Aerospace Corporation Portable Pulmonary Function System

A research project designed to investigate changes in maximal oxygen consumption (VO2max) during and following long duration flight on the International Space Station (ISS) has recently been completed. The device used to measure VO2 on board ISS, the Portable Pulmonary Function System (PPFS) manufactured by the Danish Aerospace Corporation (DAC), is based on previous-generation devices manufactured by DAC, but the PPFS has not been validated for analyzing metabolic gases or measuring cardiac output (Qc). The purpose of the present evaluation is to compare PPFS metabolic gas analysis measurements to measurements obtained using a clinically-validated system (ParvoMedics TrueOne(c) 2400 system; Parvo). In addition, Qc data collected with the PPFS were compared to Qc measurements from echocardiography. METHODS: Ten subjects completed three cycle exercise tests to maximal exertion. The first test was conducted to determine each subject's VO2max and set the work rates for the second and third (comparison) tests. The protocol for the two comparison tests consisted of three 5-minute stages designed to elicit 25%, 50%, and 75% VO2max (based upon results from the initial test), followed by 1-minute stages of increasing work rate (25 watt/minute) until the subject reached maximal effort. During one of the two comparison tests, metabolic gases and Qc were assessed with the PPFS; metabolic gases and Qc were assessed with the Parvo and by echocardiography, respectively, during the other test. The order of the comparison tests was counterbalanced. VO2max and maximal work rate during the comparison tests were compared using t tests. Mixed-effects regression modeling was used to analyze submaximal data. RESULTS: All of the data were within normal physiological ranges. The PPFS-measured values for VO2max were 6% lower than values obtained with the Parvo (PPFS: 3.11 +/- 0.75 L/min; Parvo: 3.32 +/- 0.87 L/min; mean +/- standard deviation; P = 0.02); this difference is probably due to flow restriction imposed by the PPFS Qc accessories. Submaximal VO2 values were slightly lower when measured with the PPFS, although differences were not physiologically relevant. The PPFS-measured values of submaximal carbon dioxide production (VCO2) were lower than the data obtained from Parvo, which could be attributed to lower fractions of expired carbon dioxide measured by the PPFS. The PPFS Qc values tended to be lower than echocardiography-derived values. CONCLUSIONS: The results of the present study indicate a need to further examine the PPFS and to better quantify its reproducibility; however, none of the findings of the current evaluation indicate that the PPFS needs to be replaced or modified.

Portable Pulmonary Function System (PPFS)↗

NASA NTRS2019

Vortex Radiometry: Fundamental Concepts

Vortex radiometers (VRs) enable communication systems to maximize data throughput by determining when to employ fade mitigation most efficiently. With this information, communication systems can optimize mitigation strategies before actual fades occur. Doing so maximizes the time over which the link is available, and therefore maximizes data throughput. This paper presents the fundamental concepts of vortex radiometry, including the creation of annular beam patterns through the use of orbital angular momentum (OAM). These concepts are simulated on a 100 Mbps (Megabits per second) GEO (Geosynchrous Earth Orbit)-to-ground Ka-Band communication link. During the simulation a noise source traverses across the link, and without VR capabilities the link is disrupted. However, when the situation is repeated with VR capabilities enabled, the antenna is able to increase receiver gain and maintain the link. VRs are uniquely suited to provide communication systems with the ability to optimize receiver parameters and maximize data throughput, making them critical to enabling next generation communication networks.

Orbital Angular Momentum↗

NASA NTRS2021

A sampling-based optimization approach to handling environmental uncertainty for a planetary lander

Planning for unknown environments presents a number of technical challenges. The planner must ensure robustness to unknown phenomena and manage unpredictable variation in execution, all while operating in a capacity that maximizes its objective. Productivity in the face of these challenges re-quires an integrated approach to planning and execution that is capable of accomplishing goals, reacting to variation, and maximizing overall utility. We examine this problem in the context of a Europa Lander concept mission. We model the problem as a hierarchical task network, framing it as a utility maximization problem constrained on a depletable energy resource. We propose an uncertainty–sensitive deterministic planning framework that utilizes periodic replanning to better handle model uncertainty and variable execution. We demonstrate the efficacy of our framework through simulations of a Europa Lander concept mission in which our algorithm out-performs several baseline approaches in both utility maximization and robustness

Zilberstein, Shlomo↗

NASA NTRSDate not supplied

Design and Testing of the BionicWingSat in a Zero-g Flight Campaign - A 2U-CubeSat with Deployable, Biologically-Inspired Wings

In this paper, recent developments in the design, manufacturing, and testing of a novel deployable structure with several potential applications in space will be described. Through a cooperative effort of the German Aerospace Center (DLR) and the National Aeronautics and Space Administration (NASA), a biologically inspired structurally integrated membrane featuring distributed functional elements has been developed and tested in a 2U CubeSat called BionicWingSat. Such a membrane structure could be useful for several applications in which a relatively flat area is desirable such as solar sails, drag sails, or solar shades. For SmallSats and CubeSats, the design proposed also has the desirable property of being self-deploying without the need for powered deployment mechanisms. Building on previous work inspired by the wings of earwigs, the research presented in this paper includes structural design of self-deploying hinges, a survey of various advanced additive layer manufacturing (ALM) methods for making hinges, mechanical characterization of the hinges, and finite element analysis (FEA) of the hinges. In this work, the conflicting goals of maximizing deployed structural stiffness, maximizing deployed area, maximizing stowed packaging efficiency, and maximizing resistance to creep when stowed must be considered. The resulting design concept is a gossamer structure that cannot support its own weight in gravity. For this reason, a focus in this paper is on a parabolic flight test campaign in which 24 fully integrated wings on two BionicWingSats were tested in a microgravity environment. From this test campaign, several lessons were learned regarding the wing design and procedures for carrying out microgravity tests of this manner.

CubeSat↗

NASA NTRSDate not supplied

Design and Testing of the BionicWingSat in a Zero-g Flight Campaign - A 2U-CubeSat with Deployable, Biologically-Inspired Wings

CubeSat↗

NASA NTRS2021

A Sampling-Based Optimization Approach to Handling Environmental Uncertainty for a Planetary Lander

Chien, Steve↗

DOE OSTI2020

A Hybrid Gradient Method to Designing Bayesian Experiments for Implicit Models

Bayesian experimental design (BED) aims at designing an experiment to maximize the information gathering from the collected data. The optimal design is usually achieved by maximizing the mutual information (MI) between the data and the model parameters. When the analytical expression of the MI is unavailable, e.g.,having implicit models with intractable data distributions, a neural network-based lower bound of the MI was recently proposed and a gradient ascent method was used to maximize the lower bound [1]. However, the approach in [1] requires a pathwise sampling path to compute the gradient of the MI lower bound with respect to the design variables, and such a pathwise sampling path is usually inaccessible for implicit models. In this work, we propose a hybrid gradient approach that leverages recent advances in variational MI estimator and evolution strategies (ES)combined with black-box stochastic gradient ascent (SGA) to maximize the MI lower bound. This allows the design process to be achieved through a unified scalable procedure for implicit models without sampling path gradients. Several experiments demonstrate that our approach significantly improves the scalability of BED for implicit models in high-dimensional design space.

Zhang, Jiaxin↗