Effective Throughput Research Articles

Sparse Outage Control for Wireless Internet of Things Systems Through Virtual Queue and Consecutively Effective Throughput

Sparse Outage Control for Wireless Internet of Things Systems Through Virtual Queue and Consecutively Effective Throughput

PHY Security Design for Mobile Crowd Computing in ICV Networks Based on Multi-Agent Reinforcement Learning

In this paper, we propose a multi-roadside unit (RSU) assisted mobile crowd computing framework for intelligently connected vehicle (ICV) networks, where vehicles within RSUs’ coverage act as workers to provide their computation and communication resources for computing resource limited vehicle user equipments (VUEs). Physical (PHY) layer security is used to secure computation task offloading and results feedback in time-varying vehicular channels. Artificial noise (AN) assisted adaptive wiretap coding is adopted to enhance the security of offloading links. With PHY security, the intended receiver can decode secret message while eavesdropper cannot. A modified exhaustive two-dimensional (2D) search algorithm is proposed to optimize transmission rate and secrecy rate in an effective secrecy throughput maximization problem, and a multi-agent twin delayed deep deterministic policy gradient algorithm (MATD3) is utilized to assign VUEs’ tasks without a central controller, where a reward function is defined according to the computing costs, including execution time, energy consumption, and price paid for computing. Finally, simulations verify the effectiveness of the proposed framework.

A Covert Jamming Scheme Against an Intelligent Eavesdropper in Cooperative Cognitive Radio Networks

In this work, we design a covert jamming scheme against an intelligent Eve towards physical layer security in a cooperative cognitive radio networks. To protect the primary message from being decoded by an intelligent eavesdropper (Eve), a secondary user is picked as a friendly jammer that transmits artificial noise (AN) to the Eve. However, the intelligent Eve can detect the existence of AN, and thus the AN is inclined to be covertly transmitted at a lower detection probability, thereby mistaking the AN as background noise by the Eve. In this paper, firstly, by using a binary hypothesis testing, the detection error probability (DEP) is explicitly derived for detection performance analysis of the Eve in the perfect as well as statistical channel state information (CSI) scenario, separately. With the DEP, a covert jamming scheme is formulated to transmit the AN in both cases. Then, an optimization problem is formulated, and then recast into a two-part problem to maximize the secrecy performance with the covert jamming scheme in the perfect CSI case. Finally, the effective secrecy throughput is maximized under convert constraints in the statistical CSI case. Numerical results are provided to validate the DEP at the Eve and secrecy performance analysis.

Effects of Eavesdropper on the Performance of Mixed $\eta$–$\mu$ and DGG Cooperative Relaying System

Free-space optical (FSO) channels offer line-of-sight wireless communication with high data rates utilizing unlicensed optical spectrum. The hybrid radio frequency (RF)–FSO system is currently receiving a great deal of research interest because atmospheric turbulence prevents enhanced secrecy performance. Except for the double generalized gamma (DGG) model, traditional FSO models typically are unable to show secrecy performance across all turbulence severity ranges. Due to this, we propose a dual-hop <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\eta$</tex-math></inline-formula> – <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\mu$</tex-math></inline-formula> and unified DGG mixed RF–FSO network while taking into account the following three different eavesdropping scenarios: 1) the eavesdropper is at the RF link; 2) the eavesdropper is at the FSO link; and 3) both eavesdroppers remains active simultaneously. These proposed scenarios' security is examined in terms of the probability, of strictly positive secrecy capacity, secure outage probability and effective secrecy throughput, and further investigation is done into the implications of various system settings on the secrecy performance. The analysis shows that, when compared to the other two scenarios, the third scenario exhibits the worst secrecy condition. Intensity modulation/direct detection (IM/DD) and heterodyne detection (HD) approaches are compared, with HD outperforming IM/DD in terms of secrecy performance. Finally, Monte Carlo simulations are used to confirm all analytical findings.

InferFair: Towards QoS-aware scheduling for performance isolation guarantee in heterogeneous model serving systems

With the popularity of Deep Neural Network (DNN) models in diverse fields, DNN inference services have been widely deployed on cloud for resource-limited devices to support intelligent applications. Serving DNN inference often requires GPU acceleration to meet latency-sensitive interactive targets. A common approach to improving GPU utilization is to let multiple models share a GPU. However, this approach probably degrades both the responsiveness and throughput of model serving systems, due to that concurrent DNN inference tasks contend for GPU resources. In addition, interferences among heterogeneous DNN inference tasks probably incur performance isolation problem that heterogeneous models suffer from different levels of serving performance degradation. Existing works fail to ensure performance isolation among users of heterogeneous DNN models. To solve the aforementioned problem, we propose InferFair, a QoS-aware scheduling framework for ensuring performance isolation in heterogeneous model serving systems. InferFair focuses on two key designs: (1) periodically estimating effective throughput requirements of all active models online and (2) applying fine-grained adjustments to minimize the impact differences of GPU sharing on heterogeneous model services. We conduct intensive experiments on a variety of DNN models to demonstrate the effectiveness of InferFair. Compared to a prior competitor named Clockwork, InferFair alleviates the performance isolation problem by up to 1.7×, as well as improving the overall goodput by up to 25.6%.

High throughput Luminex beads based multiplex assay for identification of six major bacterial pathogens of mastitis in dairy animals.

Bovine mastitis is caused by over 150 different microorganisms. Specific identification and quantification of multiple bacteria in a single milk sample becomes essential for rapid intervention. In the present study a Luminex beads based multiplex assay emphasizing on the precise identification of six major bacterial pathogens of mastitis was developed. Assay was developed in two triplex sets, triplex 1 comprised of Streptococcus agalactiae, Streptococcus dysgalactiae and Streptococcus uberis while triplex 2 consisted of Staphylococcus aureus, E. coli and Klebsiella pneumoniae. The analytical sensitivity was 10 6 copies per reaction mixture for all the six bacteria. A 100% analytical specificity was observed for simultaneous detection of these bacteria. Clinical milk samples from 100 bovine quarters were tested for validation. The analytical sensitivity was similar to the findings reported earlier in real time PCR multiplex assay targeting the DNA of the 11 most common bacterial species or groups in mastitis. The analytical specificity of the optimized assay was 100% similar to reported earlier for simultaneous detection of Mycoplasma spp. and for seven entric viruses of humans.The developed assay indicates a concept proof of a rapid, cost effective high throughput diagnostic tool for identification of major bacteria causing mastitis.

Large-Scale Rate-Splitting Multiple Access in Uplink UAV Networks: Effective Secrecy Throughput Maximization Under Limited Feedback Channel

Unmanned aerial vehicles (UAVs) are capable of improving the performance of next generation wireless systems. However, their communication performance is prone to both channel estimation errors and potential eavesdropping. Hence, we investigate the effective network secrecy throughput (ENST) of the uplink UAV network, in which rate-splitting multiple access (RSMA) is employed by each legitimate user for secure transmission under the scenario of massive access. To maximize the ENST, the transmission rate versus power allocation relationship is formulated as a max-min optimization problem, relying on realistic imperfect channel state information (CSI) of both the legitimate users and passive eavesdroppers ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$Eves$</tex-math></inline-formula> ). In the model considered, each user transmits a superposition of two messages to a UAV base-stations (UAV-BS), each having different transmit power and the UAV-BS adopts a successive interference cancellation (SIC) technique to decode the received messages. Given the non-convexity of the problem, it is decoupled into a pair of sub-problems. In particular, we derive a closed form expression for the optimal rate-splitting fraction of each user. Then, given the optimal rate-splitting fraction of each user, the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\epsilon$</tex-math></inline-formula> -constrainted transmit power of each user is calculated by harnessing sequential parametric convex approximation (SPCA) programming. Our simulation results confirm that the scheme conceived significantly improves the ENST compared to both the existing orthogonal and non-orthogonal benchmarks.

Joint Packet Length and Power Optimization for Covert Short-Packet D2D Communications

This paper proposes a joint optimization mechanism for packet length and power in the scenario of covert short packet D2D communication, so as to effectively improve the communication performance between D2D pairs subject to the covert constraint. Specifically, we construct a short-packet D2D communication model assisted by covert communication, and further propose the notion of effective covert throughput (ECT) to quantitatively characterize the trade-off between the covertness and reliability of IoT state monitoring information transmission. Secondly, in the constructed communication scenario, we analyze the detection error probability of the warden and clarify that the existing equal power transmission of pilot and data signals can minimize the detection performance of the warden. However, this strategy is achieved by compromising the transmission performance of the system, which means that the ECT of D2D pair may not be optimal. Thirdly, we aim to maximize the ECT of D2D pair and construct a joint optimization problem for pilot transmission power, data transmission power, and packet length. Furthermore, a joint optimization algorithm based on the 2D search is adopted to obtain the optimal solution of the established optimization problem. Simulation results demonstrated that the transmission performance of the joint optimization algorithm is better than that of the scheme of the equal power scheme on the premise of ensuring the covertness.

AdaptChain: Adaptive Scaling Blockchain With Transaction Deduplication

Although existing schemes improve blockchain throughput by allowing concurrent blocks to be appended to the blockchain, little attention has been devoted to adjusting blockchain throughput dynamically and deduplicating transactions between concurrent blocks. In this article, we propose AdaptChain, an adaptive scaling blockchain with transaction deduplication. When the transaction demand of users in the network is high, the blockchain expands to meet the demand; when the transaction demand is low, the blockchain shrinks to save communication and storage costs. Our transaction deduplication mechanism ensures that no duplicate transactions are added to the blockchain, thereby improving bandwidth utilization and achieving higher effective throughput. Besides, we randomly split the mining power of the system to achieve mining power load balancing and resist attacks. We formally analyze the blockchain security and implement the proposed prototype on Amazon EC2. Experimental results show that AdaptChain achieves dynamic and higher effective blockchain throughput.

Rapid digital pathology of H&E-stained fresh human brain specimens as an alternative to frozen biopsy

BackgroundHematoxylin and Eosin (H&E)-based frozen section (FS) pathology is presently the global standard for intraoperative tumor assessment (ITA). Preparation of frozen section is labor intensive, which might consume up-to 30 minutes, and is susceptible to freezing artifacts. An FS-alternative technique is thus necessary, which is sectioning-free, artifact-free, fast, accurate, and reliably deployable without machine learning and/or additional interpretation training.MethodsWe develop a training-free true-H&E Rapid Fresh digital-Pathology (the-RFP) technique which is 4 times faster than the conventional preparation of frozen sections. The-RFP is assisted by a mesoscale Nonlinear Optical Gigascope (mNLOG) platform with a streamlined rapid artifact-compensated 2D large-field mosaic-stitching (rac2D-LMS) approach. A sub-6-minute True-H&E Rapid whole-mount-Soft-Tissue Staining (the-RSTS) protocol is introduced for soft/frangible fresh brain specimens. The mNLOG platform utilizes third harmonic generation (THG) and two-photon excitation fluorescence (TPEF) signals from H and E dyes, respectively, to yield the-RFP images.ResultsWe demonstrate the-RFP technique on fresh excised human brain specimens. The-RFP enables optically-sectioned high-resolution 2D scanning and digital display of a 1 cm2 area in <120 seconds with 3.6 Gigapixels at a sustained effective throughput of >700 M bits/sec, with zero post-acquisition data/image processing. Training-free blind tests considering 50 normal and tumor-specific brain specimens obtained from 8 participants reveal 100% match to the respective formalin-fixed paraffin-embedded (FFPE)-biopsy outcomes.ConclusionsWe provide a digital ITA solution: the-RFP, which is potentially a fast and reliable alternative to FS-pathology. With H&E-compatibility, the-RFP eliminates color- and morphology-specific additional interpretation training for a pathologist, and the-RFP-assessed specimen can reliably undergo FFPE-biopsy confirmation.

DUPLETS: Deformability-Assisted Dual-Particle Encapsulation Via Electrically Activated Sorting.

Co-encapsulation of bead carriers and biological cells in microfluidics has become a powerful technique for various biological assays in single-cell genomics and drug screening because of its distinct capability of single-cell confinement. However, current co-encapsulation approaches exist a trade-off between cell/bead pairing rate and probability of multiple cells in individual droplets, significantly limiting the effective throughput of single-paired cell-bead droplets production. Deformability-assisted dUal-Particle encapsuLation via Electrically acTivated Sorting (DUPLETS) system is reported to overcome this problem. The DUPLETS can differentiate the encapsulated content in individual droplets and sort out targeted droplets via a combined screening of mechanical and electrical characteristics of single droplets in label-free manners and with the highest effective throughput in comparison to current commercial platforms. The DUPLETS has been demonstrated to enrich single-paired cell-bead droplets to over 80% (above eightfold higher than current co-encapsulation techniques). It eliminates multicell droplets to 0.1% whereas up to ≈24% in 10× Chromium. It is believed that merging DUPLETS into the current co-encapsulation platforms can meaningfully elevate sample quality in terms of high purity of single-paired cell-bead droplets, low fraction of multicell droplets, and high cell viability, which can benefit a multitude of biological assay applications.

Flexible Rate-Splitting Multiple Access With Finite Blocklength

Rate-splitting multiple access (RSMA) is a promising multiple access (MA) technique. It employs rate-splitting (RS) at the transmitter and successive interference cancellation (SIC) at the receiver. Most of the existing works on RSMA assume that all users use SIC to decode the common stream and the blocklength is infinite. The first assumption causes the data rate of the common stream to be limited by the user with the worst channel quality. The second assumption may lead to suboptimal performance in practical systems with finite blocklength. In this paper, we propose a flexible RSMA scheme, which allows the system to decide whether a user should use SIC to decode the common stream or not. We consider the effective throughput as the performance metric, which incorporates the data rate as well as the error performance of RSMA with finite blocklength. We first derive the effective throughput expression and then formulate an effective throughput maximization problem by jointly optimizing the beamforming vectors, transmission data rates, and RS-user selection. We develop an optimal algorithm as well as a low-complexity algorithm for beamforming design. We derive a semi-closed-form solution of the optimal data rates and propose an efficient algorithm for the RS-user selection. Numerical results demonstrate that the proposed algorithm obtains a higher effective throughput than space division multiple access (SDMA), non-orthogonal multiple access (NOMA), and two other RSMA schemes.

Deep Reinforcement Learning Enabled Covert Transmission With UAV

This paper considers covert communications in the context of unmanned aerial vehicle (UAV) networks, where a UAV is employed as a base station to transmit covert data to a legitimate ground user, while ensuring that the data transmission cannot be detected by a warden. Aiming at maximizing the legitimate user’s average effective covert throughput (AECT), the UAV’s trajectory and transmit power are jointly optimized. Taking advantage of deep reinforcement learning (DRL) on solving dynamic and unpredictable problems, we develop a twin-delayed deep deterministic policy gradient aided covert transmission algorithm (TD3-CT), to determine the UAV’s optimal trajectory and transmit power. Furthermore, by introducing a reward shaping mechanism, the convergence of the algorithm is guaranteed. The experiment results show that the developed TD3-CT algorithm not only enables the covert transmission but also significantly improves its performance in termed of achieving a higher AECT, compared with the benchmark schemes.

Data in brief: Factors associated with in-patient admission among stroke patients (P1-5.024)

<h3>Objective:</h3> To estimate the probability of admission among patients with stroke evaluated in the emergency department. <h3>Background:</h3> Misuse of inpatient resources is a central problem to the healthcare system. Elucidating certain trends behind inpatient admission in patients with a cerebral infarction may reveal opportunities to improve safe and effective stroke triage and throughput. <h3>Design/Methods:</h3> We retrospectively queried the 2019 National Emergency Department Sample Database for patients who had a cerebral infarction. Disposition from the emergency department was classified into one of several categories: routine discharge, discharge against medical advice, transfer to short term hospital, home care, death, and admission to the inpatient setting. The primary endpoint was inpatient admission versus all other dispositions and was assessed using multivariable logistic regression after adjustment for key patient characteristics, hospital teaching and trauma status, geographic regions, insurance status. <h3>Results:</h3> Of the 598,818 patients diagnosed with acute cerebral infarction, 471,464 (78.7%) were admitted to an inpatient acute care facility. Patients with private insurance (OR 0.85, 95%CI 0.78–0.93) or self-pay (OR 0.80, 95%CI 0.70–0.91) had lower odds of being admitted, compared to Medicare patients. Compared to Metropolitan Non-Teaching hospitals, Metropolitan teaching hospitals (OR 1.59, 95%CI 1.27–1.97) had a higher odds of admission, whereas non-metropolitan hospitals (OR 0.35, 95%CI 0.27–0.44) had a lower odds of admission. Additionally, compared to non-trauma hospitals, patients evaluated at level 1 trauma centers (OR 3.17, 95%CI 2.48–4.05) had higher odds of admission. In addition, hyperlipidemia (OR 2.25, 95%CI 2.09–2.42), overweight (OR 2.15, 95%CI 1.90–2.42), and hypertension (OR 1.55, 95%CI 1.47–1.63) were independently associated with a higher odds of admission. <h3>Conclusions:</h3> Optimization of certain high-risk comorbidities may not only reduce stroke incidence, but they may reduce inpatient utilization. Further, differences in admission based on patient-level socioeconomic factors warrant greater exploration to minimize disparities in access to acute care. <b>Disclosure:</b> Mr. Wolfe has nothing to disclose. Karan Patel has nothing to disclose. Mr. Taneja has nothing to disclose. Mr. Oak has nothing to disclose. The institution of Dr. Favilla has received research support from National Institutes of Health. The institution of Dr. Favilla has received research support from American Heart Association. The institution of Dr. Favilla has received research support from OpenWater, Inc.. An immediate family member of Dr. Thon has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Horizon. An immediate family member of Dr. Thon has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Genentech. An immediate family member of Dr. Thon has received personal compensation in the range of $500-$4,999 for serving on a Speakers Bureau for Genentech. Dr. Thon has received personal compensation in the range of $500-$4,999 for serving as a Focus group participant with Alexion. Dr. Siegler has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Ceribell. Dr. Siegler has received personal compensation in the range of $10,000-$49,999 for serving on a Speakers Bureau for AstraZeneca.

Effective multi‐step ahead container throughput forecasting under the complex context

AbstractAccurate and effective container throughput forecasting plays an essential role in economic dispatch and port operations, especially in the complex and uncertain context of the global Covid‐19 pandemic. In light of this, this research proposes an effective multi‐step ahead forecasting model called EWT‐TCN‐KMSE. Specifically, we initially use the empirical wavelet transform (EWT) to decompose the original container throughput series into multiple components with varying frequencies. Subsequently, the state‐of‐the‐art temporal convolutional network is utilized to predict the decomposed components individually, during which an improved loss function that combines mean square error (MSE) and kernel trick is employed. Eventually, the deduced prediction results can be obtained by integrating the predicted values of each component. In particular, this research introduces the MIMO (multi‐input and multi‐output) strategy to conduct multi‐step ahead container throughput forecasting. Based on the experiments in Shanghai port and Ningbo‐Zhoushan port, it can be found that the proposed model shows its superiority over benchmark models in terms of accuracy, stability, and significance in container throughput forecasting. Therefore, our proposed model can assist port operators in their daily management and decision making.

A simple metric that correlates with public Wi‐Fi throughput

AbstractIn the IEEE 802.11 specifications, a wireless fidelity (Wi‐Fi) access point (AP) broadcasts a beacon frame every 0.1024 s, and the received signal strength indicator (RSSI) is measured at a user terminal. RSSI is used as a metric for AP selection, but it does unlikely correlate with the actual quality of service, as has been pointed out in previous literature. In this letter, a simple metric is proposed that correlates with the effective throughput offered by a public Wi‐Fi AP. Specifically, the number of beacon frames within a time window is counted instead of using RSSI. The proposed metric can be easily calculated as it does not rely on advanced techniques such as machine learning. In a realistic IEEE 802.11a/n/ac network, the correlation coefficient between the proposed metric and the effective total throughput is 0.874 in the time window of 1 s, which is higher than a conventional metric, indicating the potential for low‐latency identification of whether a public Wi‐Fi AP can provide good throughput.

Contention or Symbiosis: The Model-Based and Model-Less Transmission Mode Selections for Internet of Things in Unlicensed Band

In this work, we investigate a spectrum-sharing network, where an Internet-of-Things (IoT) system and a WiFi system coexist in the unlicensed band. The IoT devices can actively transmit signal by contending for the channel with the WiFi users. Alternatively, they can also convey information by passively backscattering the ambient WiFi signal, and thus form symbiosis with the WiFi during the transmission. Since the active and the passive transmission mode have their own merits under specific circumstances, how to select one of them adaptively and efficiently with guaranteeing the normal service of the WiFi system is of high importance. To solve this problem, we firstly investigate the effective throughput of the IoT system from the cross-layer perspective. Then, by jointly considering the effective throughput and the power consumption of the IoT system, we formulate the transmission mode selection problem. Two solutions are given depending on whether the signaling between the two systems is available. By entailing the inter-system parameter sharing, the model-based transmission mode selection scheme is proposed, which gives the closed-form optimal solution. When the inter-system parameter sharing is absent, the model-less transmission mode selection scheme is proposed by using deep reinforcement learning (DRL) technique. Extensive simulation results validate the theoretical analysis, and demonstrate the effectiveness of the proposed transmission mode selection schemes.

Evaluating Analog Self-Interference Cancellation for In-Band Full-Duplex Wireless Communication

In-band full-duplex (IBFD) technology allows simultaneous transmission and reception in the same frequency band to boost spectral efficiency. However, selfinterference, introduced by transmitter-to-receiver leakage, poses the biggest obstacle to the realization of IBFD. Altogether IBFD radio performance and effective network throughput are directly correlated with the amount of energy suppressed by self-interference cancellation (SIC). In this work a model of an IBFD transceiver, featuring an adaptive, analog-domain, tapped-branch self-interference canceller, was developed to evaluate the inimical effects attributable to the various noise and nonlinear distortions of the simulated transceiver components. Results of the simulation show that noise incursions and nonlinearity in the transmission channel produced a negligible effect on the overall cancellation performance. However, the effectiveness of the cancellation hardware is highly dependent on the level of intrinsic nonlinearity within the canceller’s own components. Consideration of these non-ideal characteristics should be an essential part of analog SIC design to prevent any intrinsic limitations on the cancellation performance.

On Effective Secrecy Throughput of Underlay Spectrum Sharing $\alpha -\mu$/ Málaga Hybrid Model Under Interference-and-Transmit Power Constraints

The underlay cognitive radio-based hybrid radio frequency/free-space optical (RF/FSO) systems have emerged as a promising technology due to their ability to eliminate spectrum scarcity and spectrum under-utilization problems. The physical layer security of such a network with a primary user, a secondary source, a secondary receiver, and an eavesdropper is therefore examined in this work. In this network, secret communication occurs between two reliable secondary peers over the RF and FSO links simultaneously, and the eavesdropper can only overhear the RF link. In particular, the maximum transmits power limitation at the secondary user as well as the permissible interference power restriction at the primary user are also taken into consideration. All the RF and FSO links are modeled with <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\alpha$</tex-math></inline-formula> - <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\mu$</tex-math></inline-formula> fading and Málaga turbulence with link blockage and pointing error impairments. At the receiver, the selection combining diversity technique is utilized to select the signal with the best electrical signal-to-ratio (SNR). Furthermore, to examine the secrecy performance taking into account the effects of each system parameter, closed-form expressions for the secrecy outage probability and effective secrecy throughput are derived. The resultant expressions are finally verified by Monte-Carlo simulations.

AdaChain: A Learned Adaptive Blockchain

This paper presents AdaChain , a learning-based blockchain framework that adaptively chooses the best permissioned blockchain architecture to optimize effective throughput for dynamic transaction workloads. AdaChain addresses the challenge in Blockchain-as-a-Service (BaaS) environments, where a large variety of possible smart contracts are deployed with different workload characteristics. AdaChain supports automatically adapting to an underlying, dynamically changing workload through the use of reinforcement learning. When a promising architecture is identified, AdaChain switches from the current architecture to the promising one at runtime in a secure and correct manner. Experimentally, we show that AdaChain can converge quickly to optimal architectures under changing workloads and significantly outperform fixed architectures in terms of the number of successfully committed transactions, all while incurring low additional overhead.