Online Optimization Research Articles

“Store Strategy”: A New Omni‐Channel Strategy in Community Group Buying

ABSTRACTThis paper explores a new omni‐channel strategy—“store strategy”—in community group buying and investigates the impacts of the application of this strategy on the platform and community leader. Three models are developed: (1) The platform operates only in a single online channel. (2) The platform adopts a store strategy and delegates offline operations to a community leader. (3) The platform adopts a store strategy that includes self‐establishing a store. After comparing different models, some interesting results are obtained as follows: (i) After the store strategy is adopted, the optimal online price remains unchanged, but new offline pricing varies, depending on the offline entry method. (ii) If the cost of the community leader operating the offline channel is low, the proportion of private consumers is particularly high. In this case, the store strategy that delegates operations to the community leader will achieve a win–win outcome. (iii) If the self‐established cost is low and the proportion of private consumers is low, the store strategy with a self‐establishing store will allow the platform to obtain higher profits, while greatly harming the community leader.

Robotic Warehousing Operations: A Learn-Then-Optimize Approach to Large-Scale Neighborhood Search

The rapid deployment of robotics technologies requires dedicated optimization algorithms to manage large fleets of autonomous agents. This paper supports robotic parts-to-picker operations in warehousing by optimizing order–workstation assignments, item–pod assignments, and the schedule of order fulfillment at workstations. The model maximizes throughput, managing human workload at the workstations and congestion in the facility. We solve it via large-scale neighborhood search with a novel learn-then-optimize approach to subproblem generation. The algorithm relies on an off-line machine learning procedure to predict objective improvements based on subproblem features and an online optimization model to generate a new subproblem at each iteration. In collaboration with Amazon Robotics, we show that our model and algorithm generate much stronger solutions for practical problems than state-of-the-art approaches. In particular, our solution enhances the utilization of robotic fleets by coordinating robotic tasks for human operators to pick multiple items at once and by coordinating robotic routes to avoid congestion in the facility. Funding: This research was partially funded by Amazon.com Services LLC under award number 2D-12552417. Supplemental Material: The online appendix is available at https://doi.org/10.1287/ijoo.2024.0033 .

Strengthening social presence and connectedness in 100% online courses through CONNECT

Over recent years, higher education has seen a significant shift toward online courses. Factors driving this change include digital technology ubiquity, diverse student populations, workforce demands, and the increasing need for more flexible learning environments (Serrano et al., 2019). However, while online course enrolments are rising faster than on-campus courses, retention rates remain lower (Australian Government, 2021). Evidence suggests that online education can negatively impact students’ sense of connection, leading to feelings of isolation and disempowerment (Roddy et al., 2017). This is also reflected in the Australian University Mental Health framework (Orygen, 2020) which emphasises creating supportive online learning environments that promote student well-being. The Australian Universities Accord (Australian Government, 2024) calls for universities to embrace TEL. However, given that research has shown the need for more professional development (PD) and support for online educators (Ahern and Biedermann, 2022; 2023), innovative, evidence-based strategies are needed for organisations to build capacity in online teaching. Drawing on a comprehensive review of literature, the authors developed CONNECT: a framework to enhance social presence and connectedness in online learning (Ahern et al., 2024). Grounded in research underpinned by Garrison et al.’s (2000) Community of Inquiry (CoI) framework, the principles of Universal Design for Learning (CAST, 2018) and leveraging TEL, this framework offers educators a roadmap for nurturing a sense of community that positively influences student well-being, engagement, and overall success. Key components of the CONNECT framework are: C - Communication is responsive and supportive O - Optimal online presence N - Nurture educational and professional growth N - Networking opportunities fostered E - Engaging communities of learning C - Contemporary and authentic content T - Timely and effective feedback for learning (Ahern et al., 2024). The study follows a quasi-experimental design to evaluate the effectiveness of CONNECT in guiding teaching staff to use evidence-based strategies to promote belonging and connectedness in online education. In phase one, CONNECT was implemented within a 100% online postgraduate (PG) subject. This presentation will showcase online teaching strategies used to apply all elements of CONNECT as implemented in phase one. Phase two, also showcased, involved creating and releasing an online short course as professional development for teaching staff across two PG online courses at the authors’ university: Master of Nursing (MN) and Graduate Certificate of Diabetes Education (GCDE). After these implementation phases, the third phase will measure the effectiveness of CONNECT across the two PG courses. Data will be collected via (i) a student survey and (ii) a staff survey. Data analysis will compare the experimental group (participants enrolled or teaching in MN or GCDE) against the control group (participants not enrolled or teaching in MN or GCDE). Data will be collected from students and teachers across Australian universities offering 100% online courses. We hypothesise that by implementing CONNECT, the study will (1) Show improved student satisfaction in the experimental group and (2) Add to the body of evidence to inform and support teaching staff to develop strategies that can help build connectedness and social presence in online learning.

Clipper: Online Joint Client Sampling and Power Allocation for Wireless Federated Learning

Communication overhead is a main bottleneck in federated learning (FL) especially in the wireless environment due to the limited data rate and unstable radio channels. The communication challenge necessitates holistic selection of participating clients that accounts for both the computation needs and communication cost, as well as judicious allocation of the limited transmission resource. Meanwhile, the random unpredictable nature of both the training data samples and the communication channels requires an online optimization approach that adapts to the changing system state over time. In this work, we consider a general framework of online joint client sampling and power allocation for wireless FL under time-varying communication channels. We formulate it as a stochastic network optimization problem that admits a Lyapunov-typed solution approach. This leads to per-training-round subproblems with a special bi-convex structure, which we leverage to propose globally optimal solutions, culminating in a meta algorithm that provides strong performance guarantees. We further study three specific FL problems covering multiple scenarios, namely with IID or non-IID data, whether robustness against data drift is required, and with unbiased or biased client sampling. We derive detailed algorithms for each of these problems. Simulation with standard classification tasks demonstrate that the proposed communication-aware algorithms outperform their counterparts under a wide range of learning and communication scenarios.

Integrated Hybrid Modelling and Surrogate Model-Based Operation Optimization of Fluid Catalytic Cracking Process

Fluid Catalytic Cracking (FCC) is one of the most important conversion processes in oil refineries, widely used to convert high-boiling, high-molecular-weight hydrocarbon components from crude oil into more valuable products like gasoline and diesel. Advanced simulation and optimization technologies are critical for improving the operational efficiency and economic performance of the FCC process. First-principles-based simulators rely on parameter estimation and are computationally intensive, making them unsuitable for online optimization. In recent years, with the development of deep learning, data-driven models have made significant progress in FCC modeling. However, due to their black-box nature and difficulty with extrapolation, they are rarely used for optimization. To bridge this gap, we propose an integrated framework that combines hybrid modeling and surrogate model-based optimization. This approach combines plant and simulation data to train a multi-task learning prediction model, which then serves as a surrogate for operational optimization. Validated on a large-scale FCC unit in southern China, the model predicts product yields with an error margin of under 4.84% for all products. Following optimization, yields of LNG, gasoline, and diesel rose by an average of 0.10 wt%, 1.58 wt%, and 1.05 wt%, respectively, resulting in a 3.67% increase in product revenues. This highlights the substantial potential of this framework for industrial applications.

Research on online optimization scheme and deployment of PMSM control parameters based on honey badger algorithm

Permanent magnet synchronous motor (PMSM) drive systems are receiving increasing attention due to their superior control quality and efficiency. Optimizing the control parameters are key to achieve the high-performance operation of PMSMs. Although heuristic algorithms demonstrate excellent optimization outcomes in simulations, there are still challenges in deploying optimization schemes in practical drives. In this study, a real-time online deployable control parameter optimization scheme is proposed. The optimization effect is evaluated through the system step response performance, and a framework for deploying optimization algorithms within the driver is developed. A fault suppression mechanism is also designed to mitigate overshoot and vibration issues caused by suboptimal solutions. The proposed scheme is validated on a rapid prototyping control platform. Experimental results confirm that the scheme exhibits good optimization performances across various operating conditions. The honey badger algorithm employed in this paper shows faster convergence and more stable optimization effects than other optimization algorithms. The optimization effect is improved by 2.2% and its performance in terms of consistency across multiple optimization results has increased by 40%.

R-DOCO: Resilient Distributed Online Convex Optimization Against Adversarial Attacks

This paper addresses the problem of distributed constrained optimization in a multi-agent system where some agents may deviate from the prescribed update rules due to failures or malicious adversarial attacks. The objective is to minimize the collective cost of the unattacked agents while respecting the constraint limitations. To tackle this, we propose a resilient distributed projected gradient descent algorithm for online optimization that achieves sublinear individual regret, defined as the difference between the online and offline solutions. Additionally, we extend the cost function from convex combinations to more general distributed optimization scenarios. The proposed algorithm demonstrates resilience under adversarial conditions, allowing it to handle an unknown number of adversarial nodes while maintaining performance. Compared to existing methods, this approach offers a robust solution to adversarial attacks in constrained distributed optimization problems.

An Online Train-Grid Integration Energy Optimization Method for Urban Rail System

An Online Train-Grid Integration Energy Optimization Method for Urban Rail System

Privacy Preserving Distributed Bandit Residual Feedback Online Optimization Over Time-Varying Unbalanced Graphs

Privacy Preserving Distributed Bandit Residual Feedback Online Optimization Over Time-Varying Unbalanced Graphs

ROI constrained optimal online allocation in sponsored search

Sponsored search plays a major role in the revenue contribution of e-commerce platforms. Advertising systems are designed to maximize platform revenue, but other goals also need to be considered, such as user experience, advertiser utility, and how to achieve the long-term revenue goal. A key component of a sponsored search system is online allocation, which makes real-time decisions to match users’ search requests with relevant ad campaigns to maximize platform revenue within constraints such as campaign budgets. Although much progress has been made, most of the research work on allocation problem has focused on satisfying guaranteed deals for display ads, and those challenges for allocation problems in sponsored search are not properly addressed. In this paper, we develop a framework to solve the large-scale sponsored search ad allocation problem, consisting of two main parts. One is an optimization problem solved offline by a parameter-server based architecture, and the other is an online strategy to alleviate the conflict with the auction mechanism during online service. Comprehensive offline evaluation on real production data and online A/B testing on real production system have been made. The experimental results demonstrate that through better allocating user queries to appropriate ads, the proposed model can significantly increase the platform’s revenue without sacrificing advertisers’ ROI.

Gust load alleviation of a flexible flying wing with linear parameter-varying modeling and model predictive control

This paper presents a practical model predictive control (MPC) framework for gust load alleviation of a flexible flying wing. Both the controller solving and state estimation are based on a reduced-order model, which features a linear parameter-varying (LPV) form, avoiding online linearization and reducing the scale of the corresponding quadratic programming problem. An improved modeling and model reduction process is used to enhance modeling efficiency and ensure that the reduced-order model can accurately capture the rigid-flexible coupled characteristics of the flexible flying wing under arbitrary gusts. By reconstructing the output of the control-oriented model to include both rigid-body motion and flexible vibrations, the rigid-flexible coupled multi-objective control is established as an MPC problem for reference tracking. The online optimization is formulated in a sparse fashion and combined with an iterative algorithm based on predicted trajectories, describing the variation of model dynamics within the prediction horizon more accurately. With a time-varying Kalman estimator for state updating, the closed-loop simulations are performed for gust alleviation performance validation. Additionally, the real-time potential of the proposed MPC framework is demonstrated through Monte Carlo simulations.

Distributed cooperative guidance strategy based on virtual negotiation and rolling optimization

The problem of multi-UAV cooperative guidance under target maneuvering conditions is studied, and a distributed cooperative guidance strategy based on virtual negotiation mechanism and rolling horizon optimization is proposed. Firstly, in order to improve the time coordination and guidance control effect at the same time, local coordination variables and objective functions are designed under the rolling horizon optimization framework, and a multi-constrained cooperative guidance optimization model in a limited time domain is constructed; secondly, in order to enhance the robustness of the guidance system to maneuvering targets, a disturbance observer based on nonlinear autoregressive neural network is designed; thirdly, for the control input coupling between UAVs, a virtual negotiation mechanism is proposed to achieve synchronous decision-making and eliminate internal "differences", and the root mean square propagation algorithm of Nesterov momentum is used to design the guidance command generation strategy; finally, digital simulation and semi-physical simulation experiments are carried out respectively. The results show that the proposed guidance strategy can cope with various interferences and improve the cooperative guidance effect through online prediction and optimization, and is expected to be further applied and tested in actual tasks.

Efficient nonlinear model predictive and active disturbance rejection control for trajectory tracking of unmanned vehicles

This article proposes an efficient trajectory tracking control strategy of unmanned vehicles. The method is based on nonlinear model predictive control (NMPC) and active disturbance reject control (ADRC). The designed control algorithm considers three challenges including nonlinear characteristics, multiple constraints, and external disturbance. First, NMPC method is presented for the nonlinear vehicle model with multiple constraints. To relax inequality constraints and reduce the heavy calculation burden, the penalty term with the variable factor is added to the cost function, an improved continuous/generalized minimum residual method is proposed to solve NMPC online optimization problem. Then, an ADRC scheme is designed to estimate the unknown disturbance via extended state observer, and compensate them by feedback control law in real time, the corresponding parameters are obtained by a novel particle swarm optimization algorithm to further improve the control precision. It ensures the stability to a certain extent. Finally, the results indicate the designed algorithm can raise the calculation efficiency and meet the real-time requirements, and obviously increase the tracking accuracy and robustness performance.

Development of AI-assisted microscopy frameworks through realistic simulation with pySTED

The integration of artificial intelligence into microscopy systems significantly enhances performance, optimizing both image acquisition and analysis phases. Development of artificial intelligence-assisted super-resolution microscopy is often limited by access to large biological datasets, as well as by difficulties to benchmark and compare approaches on heterogeneous samples. We demonstrate the benefits of a realistic stimulated emission depletion microscopy simulation platform, pySTED, for the development and deployment of artificial intelligence strategies for super-resolution microscopy. pySTED integrates theoretically and empirically validated models for photobleaching and point spread function generation in stimulated emission depletion microscopy, as well as simulating realistic point-scanning dynamics and using a deep learning model to replicate the underlying structures of real images. This simulation environment can be used for data augmentation to train deep neural networks, for the development of online optimization strategies and to train reinforcement learning models. Using pySTED as a training environment allows the reinforcement learning models to bridge the gap between simulation and reality, as showcased by its successful deployment on a real microscope system without fine tuning.

Model predictive control with real-time variable weight for civil aircraft towing taxi-out control systems

Unlike the traditional method of taxying, where civil aircraft reach the runway by relying on their engines, the new mode of towing taxi-out may become preferred because of its lower energy consumption, lower emissions, and higher efficiency. To improve the tracking accuracy and lateral stability of the towing system, this study used a two-level time-varying weight control method (RMPC) based on model predictive control (MPC) and fuzzy control. The tractor speed and front wheel angle were set as the control variables in the MPC controller, and the real-time lateral displacement error and yaw angle error of the tractor were set as the fuzzy control inputs. The influence of the weight matrix on the accuracy and stability of path tracking was coordinated by online optimization of the MPC objective function weight. The simulations of multiple working conditions were performed using TruckSim software in Simulink, which showed that the proposed control method can limit the lateral displacement error of the tractor and aircraft within 0.4 m, which can be reduced by 70.83% and 77.4% compared with the single MPC, respectively. Additionally, the maximum yaw angle errors of the tractor and aircraft are reduced by 76.11% and 51.31% compared with the single MPC, respectively. Furthermore, the yaw angle error of tractors can be limited to 4° and that of aircraft can be limited to 6.5°, which is an improvement of the lateral stability and driving safety for the civil aircraft towing system. The new controller may provide technical insights and support for the practical development and safe application of the towing taxi-out mode.

AI-based security functions co optimization in the SFC

Dealing with large-scale attack traffic and complex attack scenarios can be challenging for a single attack detection system in the 6G era. The ubiquitous artificial intelligence security services enabled by the AI-based Security Functions (AISF) and Service Function Chain (SFC) become strong candidate to solve this problem. However, supervised learning-based AISF requires a significant amount of manually labeled data and is unable to adapt to changing attack scenarios once it is deployed. To this end, we propose an AISF co-optimization method in the SFC. The goal is to use unlabeled network traffic samples to update the model based on pseudo-labeling and co-training. First, we model the AISF chain composed of AISFs with various detection targets and feature subspaces. We also define its detection capability evaluation metrics and final detection result. Then, we design an AISF co-optimization flow including online detection and online optimization workflows. In online optimization, we design the dynamic threshold of normalized comprehensive confidence to generate pseudo-labels for network traffic samples and combine labeled and pseudo-labeled samples to train the new models of AISFs. These models replace the previous ones and continue to detect and optimize. Experimental results in a prototype system show that compared with a single AISF, the AISF chain has higher detection capabilities that can even detect unknown attacks to a certain extent, and co-optimization can make better use of unlabeled network traffic samples than individual optimization.

Optimalisasi Dashboard Pemesanan Makanan Online Menggunakan Looker dan JavaScript

Advances in information technology have enabled rapid development in internet-based services, including online food ordering applications. This growth demands efficient data management and analysis systems to improve user experience and operational performance. This research focuses on developing an optimal online food ordering dashboard using Looker and JavaScript. Research methods include needs identification, literature study, needs analysis, design, and implementation. The research results show that the dashboard developed is able to manage and analyze order data effectively, identify trends, predict customer needs, and increase operational efficiency. This dashboard visualizes important information such as number of orders based on age, gender, income, as well as customer behavior analysis. In doing so, service providers can gain better insight into consumer behavior and ordering trends, supporting more informed and strategic decision making. The results of this study contribute to the literature on the use of data visualization technologies in the online food service sector.

Online control parameter optimization design for multi-machine coordinated loading system of hazardous substances

To address the parameter instability issues in hazardous materials handling during multi-machine loading and unloading operations, we propose a Full-Scale Smart Parameter Optimization Control (FSPOC) system specifically designed for multi-machine coordination. This system leverages a novel fish scale prediction algorithm tailored for cooperative multi-machine environments. Initially, the fish scale prediction algorithm, inspired by bionic fish scales, is developed to predict future system behavior by analyzing historical data. Building on this algorithm, we introduce a disturbance cancellation control theorem and design a parameter optimization controller to enhance stability in high-dimensional nonlinear spaces. The FSPOC method is then applied to a multi-machine cooperative system, enabling online distributed parameter optimization for complex systems with multiple degrees of freedom. The effectiveness of the proposed method was validated through simulations, where it was compared with two other optimization techniques: Genetic Algorithm-based PID (GAPID) and Chaotic Atomic Search Algorithm-based PID (CHASO). The simulation results confirm the superiority of the FSPOC method.

Formulating iron ore pellet induration process in an industrial straight grate system: Real-time experimentation and validation

Understanding the complex process of iron ore pellet induration is crucial in a straight grate system in the steelmaking industry. There is no attempt made so far to predict the behavior of this unit considering the physiochemical processes (drying, carbon combustion and limestone calcination) along with the heat transfer within the pellet, and between the hot gas and pellet. Moreover, the reported models are validated mainly with the pot test data. To address these research gaps, in this contribution, a rigorous model framework is proposed for an industrial induration unit considering all these above-mentioned issues together with some other practical aspects, including the heat transfer from hot gas to grate bar and heat loss due to air leakage into the system. To validate this rigorous formulation, attempt is further made to perform the real-time experimentation with induration system equipped with a thermocar. It is shown first time that the predicted temperature profiles of the pellets, grate bar and exit gas associated with the running bed are consistent with the plant data. The fuel consumption data is further obtained to investigate the performance of the proposed formulation at different operating regimes. It is observed that the mean absolute percentage error (MAPE) between the measured and predicted fuel rates is reasonably low (i.e., 4.2%). With this, it is recommended to use the proposed formulation for online property prediction, process design, optimization, troubleshooting, control and scale-up of the induration unit.

Lyapunov-based robust model predictive tracking controller design for discrete-time linear systems with ellipsoidal terminal constraint

In this article, a robust model predictive control (MPC) method is introduced based on Lyapunov-theory to control the discrete-time uncertain linear systems due to external disturbances. The structure of the proposed controller consists of two parts designed based on the offline/online schemes. In the offline-scheme, an H ∞ -based robust tracking controller is provided to satisfy the robust and tracking performance. Based on the H ∞ performance, a new linear matrix inequality problem is developed to calculate the offline-controller gains. By considering the Lyapunov-function of the offline-controller as the terminal cost of the MPC and the designed offline-controller, the online optimisation problem (OOP) is structured. This means, the MPC is designed based on the stabilised system to satisfy the physical limitations of the overall controller and the system states. Moreover, to improve the feasibility problem of the OOP, an ellipsoidal terminal constraint is added to the proposed MPC problem. Therefore, compared with the existing min–max MPC and tube MPC, the advantages of the overall controller are feasibility improvement and reducing the computational complexity with less conservatism while the robustness against parametric uncertainties and disturbances is achieved. Two simulation examples are employed to show the superiorities of the proposed robust MPC.