Optimal Online Algorithm Research Articles

JDAPCOO: Resource Scheduling and Energy Efficiency Optimization in 5G and Satellite Converged Networks for Power Transmission and Distribution Scenarios.

Along with the continuous revolution of energy production and energy consumption structures, the information data of smart grids have exploded, and effective solutions are urgently needed to solve the problem of power devices resource scheduling and energy efficiency optimization. In this paper, we propose a fifth generation (5G) and satellite converged network architecture for power transmission and distribution scenarios, where power transmission and distribution devices (PDs) can choose to forward power data to a cloud server data center via ground networks or space-based networks for power grid regulation and control. We propose a Joint Device Association and Power Control Online Optimization (JDAPCOO) algorithm to maximize the long-term system energy efficiency while guaranteeing the minimum transmission rate requirement of PDs. Since the formulated issue is a mixed integer nonconvex optimization problem with high complexity, we decompose the original problem into two subproblems, i.e., device association and power control, which are solved using a genetic algorithm and improved simulated annealing algorithm, respectively. Numerical simulation results show that when the number of PDs is 50, the proposed algorithm can improve the system energy efficiency by 105%, 545.05% and 835.26%, respectively, compared with the equal power allocation algorithm, random power allocation algorithm and random device association algorithm.

Distributed online adaptive subgradient optimization with dynamic bound of learning rate over time‐varying networks

Adaptive online optimization algorithms, such as Adam, RMSprop, and AdaBound, have recently been tremendously popular as they have been widely applied to address the issues in the field of deep learning. Despite their prevalence and prosperity, however, it is rare to investigate the distributed versions of these adaptive online algorithms. To fill the gap, a distributed online adaptive subgradient learning algorithm over time-varying networks, called DAdaxBound, which exponentially accumulates long-term past gradient information and possesses dynamic bounds of learning rates under learning rate clipping is developed. Then, the dynamic regret bound of DAdaxBound on convex and potentially nonsmooth objective functions is theoretically analysed. Finally, numerical experiments are carried out to assess the effectiveness of DAdaxBound on different datasets. The experimental results demonstrate that DAdaxBound compares favourably to other competing distributed online optimization algorithms.

Optimal Control for Unmanned Systems With One-Way Broadcast Communication

Unmanned systems (USs) including unmanned aerial vehicles, unmanned underwater vehicles, and unmanned ground vehicles have great application prospects in military and civil fields, among which the process of finding feasible and optimal paths for the agents in USs is a kernel problem. Traditional path finding algorithms are hard to adequately obtain optimal paths in real-time under fast time-varying and poor communication environments. We propose an online optimal control algorithm for USs based on a one-way broadcast communication mode under the assumption of a poor communication environment, mobile targets, radars (or sonar), and missiles (or torpedoes). With the principle of receding horizon control, optimal (or suboptimal) paths are then generated by the approximation theory of neural networks and gradient optimization techniques, with low computation requirements. Also, we give a convergence analysis for our algorithm, and show that each agent can reach its target in finite time under some conditions on agents, targets and radar-missiles. Moreover, simulations demonstrate that the agents in USs can generate optimal (or suboptimal) paths in real time using our algorithm while effectively avoiding collision with other agents or detection by enemy radars.

Selective Edge Computing for Mobile Analytics

An increasing number of mobile applications rely on Machine Learning (ML) routines for analyzing data. Executing such tasks at the user devices saves the energy spent on transmitting and processing large data volumes at distant cloud-deployed servers. However, due to memory and computing limitations, the devices often cannot support the required resource-intensive routines and fail to accurately execute such tasks. In this work, we address the problem of edge-assisted analytics in resource-constrained systems by proposing and evaluating a rigorous selective offloading framework. The devices execute their tasks locally and outsource them to cloudlet servers only when they predict a significant performance improvement. We consider the practical scenario where the offloading gains and resource costs are time-varying; and propose an online optimization algorithm that maximizes the service performance without requiring to know this information. Our approach relies on an approximate dual subgradient method combined with a primal-averaging scheme, and works under minimal assumptions about the system stochasticity. We fully implement the proposed algorithm in a wireless testbed and evaluate its performance using a state-of-the-art image recognition application, finding significant performance gains and cost savings.

Online Personalized Assortment Optimization with High-Dimensional Customer Contextual Data

Problem definition: Consider an online personalized assortment optimization problem in which customers arrive sequentially and make their decisions (e.g., click an ad, purchase a product) following the multinomial logit choice model with unknown parameters. Utilizing a customer’s personal information that is high-dimensional, the firm selects an assortment tailored for each individual customer’s preference. Academic/practical relevance: High dimensionality of a customer’s contextual information is prevalent in real applications, and it creates tremendous computational challenge in online personalized optimization. Methodology: In this paper, an efficient learning algorithm is developed to tackle the computational complexity issue while maintaining satisfactory performance. The algorithm first applies a random projection for dimension reduction and incorporates an online convex optimization procedure for parameter estimation, thus overcoming the issue of linearly increasing computational requirement as data accumulates. Then, it integrates the upper confidence bound method to balance the exploration and revenue exploitation. Results: The theoretical performance of the algorithm in terms of regret is derived under some plausible sparsity assumption on personal information that is observed in real data, and numerical experiments using both synthetic data and a real data set from Yahoo! show that the algorithm performs very well, having scalability and significant advantage in computational time compared with benchmark methods. Managerial implications: Our findings suggest that practitioners should process high-dimensional sparse customer data with an appropriate feature engineering technique, such as random projection (instead of abandoning the sparse portion) to maximize the effectiveness of online optimization algorithms.

Online Enhanced Semantic Hashing: Towards Effective and Efficient Retrieval for Streaming Multi-Modal Data

With the vigorous development of multimedia equipments and applications, efficient retrieval of large-scale multi-modal data has become a trendy research topic. Thereinto, hashing has become a prevalent choice due to its retrieval efficiency and low storage cost. Although multi-modal hashing has drawn lots of attention in recent years, there still remain some problems. The first point is that existing methods are mainly designed in batch mode and not able to efficiently handle streaming multi-modal data. The second point is that all existing online multi-modal hashing methods fail to effectively handle unseen new classes which come continuously with streaming data chunks. In this paper, we propose a new model, termed Online enhAnced SemantIc haShing (OASIS). We design novel semantic-enhanced representation for data, which could help handle the new coming classes, and thereby construct the enhanced semantic objective function. An efficient and effective discrete online optimization algorithm is further proposed for OASIS. Extensive experiments show that our method can exceed the state-of-the-art models. For good reproducibility and benefiting the community, our code and data are already publicly available.

An optimal online algorithm for single-processor scheduling problem with learning effect

This paper deals with the canonical single-processor online scheduling problem with the position-based learning effect. Specially speaking, a round of jobs arriving online over time will be processed on a single processor. Noticeably, in this model, for each job Jk, the actual processing time pkl is defined as a power function of its position l, i.e., pkl=pklβ, where pk indicates its normal processing time and β≤0 is the learning index. Our goal is to make the sum of completion times as small as possible. For this problem, we testify that there is no online algorithm with a competitive ratio of less than 2. Most notably, we design an online algorithm entitled as Delayed Shortest Normal Processing Time (DSNPT), matching the lower bound proposed by us, and hence DSNPT is optimal.

Recursive Set-Membership State Estimation Over a FlexRay Network

In this article, we investigate the set-membership state estimation problem for a class of time-varying systems with non-Gaussian noises over a FlexRay network. To mitigate the communication load and improve the flexibility of the data scheduling, the FlexRay protocol (FRP) governed by both the time-triggered and event-triggered rules is exploited to regulate the signal transmission in a cyclic fashion. A new expression of the input signal to the state estimator is formulated with intention to account for the effect of the FRP. Accordingly, a multirate model (orchestrating the sampling/updating rates of the target plant, sensors, and state estimator) is proposed and then transformed into a single-rate one with the help of the lifting technique and the vector augmentation method. Subsequently, sufficient conditions are provided for the true states to always reside in an ellipsoid at each time instant in the presence of the non-Gaussian noises, and such an ellipsoid is then minimized in the matrix-trace sense. An online optimization algorithm is developed to parameterize the estimator gains by means of the solution to certain recursive matrix inequalities. Numerical results demonstrate the validity of the proposed protocol-based set-membership state estimator design scheme.

Online Algorithms with Advice for the k-search Problem

In the online search problem, a seller seeks to find the maximum price from a sequence of prices 𝒑1, 𝒑2 ,..., 𝒑𝒏 that is revealed in a piece-wise manner. The bound for all prices is well known in advance with 𝒎 � 𝒑𝒊 � 𝑴. In the online 𝒌-search problem, the seller seeks to find the 𝒌 maximum out of the 𝒏 prices. In this paper, we present a tight bound of loglog22(�(���)) on the advice complexity of optimal online algorithms for online 𝒌-search. We also provide online algorithms with advice that use less than the required number of bits and compute the performance guarantee. Although it is natural to expect improvement due to the additional power of advice, we are interested to identify the relationship of additional information with respect to the improvement. We show that with 1 bit of advice, we can already surpass the quality of the best possible deterministic algorithm for online 2-search. We also provide a set of online algorithms, ALG𝒊, that utilizes loglog22(�(�𝑖𝑖 )) advice bits with a competitive ratio of 𝑖𝑀+�⁻� ⁺ ¹ (𝑘 − 𝑖)⁽�⁻� ⁾𝑚) 𝑘 𝑀 . We show that increasing the amount of advice improves the solution quality of the algorithm. Moreover, we compare the power of advice and randomization. We show that for some identified minimum number of advice bits, the lower bound on the competitive ratio of online algorithms with advice is better than any deterministic and randomized algorithm for online 𝒌-search.

MIMO architecture for fast convergence of distributed online optimization in smart grids

This paper proposes an architectural solution to enhance the resilience of distributed decision-making and control in microgrids and smart grids with respect to communication delays. The paper develops a resource allocation framework to optimally place the multiple-input multiple-output (MIMO) communication technology on critical channels to reduce the communication bottle neck and expedite the convergence speed of distributed online optimization algorithms, which require convergence in the cyber network before the solution can be implemented on the physical grid. The paper shows that optimal placement of the MIMO technology can decrease communication transmission delays, which in turn improves the reliability of power system operational protocols, such as distributed frequency regulation. Illustrative case studies using data from real-world power systems show that only upgrading the critical communication channels is sufficient to achieve system-wide performance improvement.

Joint Dynamic Task Offloading and Resource Scheduling for WPT Enabled Space-Air-Ground Power Internet of Things

The Space-Air-Ground Power Internet of Things (SAG-PIoT) can meet the communication needs of Power Internet of Things (PIoT) devices in localities with insufficient ground base station coverage. Considering that the battery capacity of PIoT devices is limited and difficult to be replaced, we use unmanned aerial vehicles (UAVs) to assist wireless power transmission (WPT) in the SAG-PIoT to ensure the network service quality and safe and stable operation of PIoT. In this paper, we put forward a novel WPT enabled SAG-PIoT architecture, based on which we investigate the dynamic task offloading and resource scheduling, and propose a joint online optimization algorithm for task assignment, local computing resource allocation, association control and UAV computing resource allocation to minimize the long-term time-averaged network operation cost. Since the coupling of task offloading and resource scheduling in the long-term stochastic optimization problem, it is decoupled into three subproblems to be solved separately using Lyapunov optimization. Moreover, we analyze the stability and sustainability of the proposed algorithm and reveal the tradeoff between the network stability and optimal network operation cost. Simulations show that the proposed algorithm has advantageous performances with respect of network stability, energy consumption and network operation cost compared to other benchmark algorithms.

Fast Online Optimization for Terrain-Blind Bipedal Robot Walking With a Decoupled Actuated SLIP Model.

We present an online optimization algorithm which enables bipedal robots to blindly walk over various kinds of uneven terrains while resisting pushes. The proposed optimization algorithm performs high-level motion planning of footstep locations and center-of-mass height variations using the decoupled actuated spring-loaded inverted pendulum (aSLIP) model. The decoupled aSLIP model simplifies the original aSLIP with linear inverted pendulum (LIP) dynamics in horizontal states and spring dynamics in the vertical state. The motion planning can be formulated as a discrete-time model predictive control (MPC) problem and solved at a frequency of 1 kHz. The output of the motion planner is fed into an inverse-dynamics–based whole body controller for execution on the robot. A key result of this controller is that the feet of the robot are compliant, which further extends the robot’s ability to be robust to unobserved terrain variations. We evaluate our method in simulation with the bipedal robot SLIDER. The results show that the robot can blindly walk over various uneven terrains including slopes, wave fields, and stairs. It can also resist pushes of up to 40 N for a duration of 0.1 s while walking on uneven terrains.

Design and Implementation of a Constrained Model Predictive Control Approach for Unmanned Aerial Vehicles

This paper designs and implements a robust and Time-varying Constrained Model Predictive Controller (TCMPC) for the translational and attitude control of a real quadrotor. The proposed approach considers online optimization to find solutions through the hard and soft constraints. All the controller parameters were derived from the experimental test setup and took into consideration the various restrictions and physical constraints associated with the hand-made quadrotor. The proposed controller can possibly linearize and discretize the nonlinear dynamic model of the quadrotor at every sampling time if all constraints and physical restrictions are considered. The performance of the proposed approach was assessed using both a simulation study and a practical implementation. The simulation study considered a quadrotor hovering mode in the presence of wind gusts and encompassed a comparison analysis with a well-tuned Proportional-Integral-Derivative (PID) controller, an Advanced Error model predictive control (AEMPC), and an Efficient MPC (EMPC) approach.. For the real-time implementation, an online optimization algorithm was used and tested on the high clock processor ARM A53 on a new attitude test setup. The experimental results, showed that the proposed controller outperformed the unconstrained MPC, the well-tuned PID controller, and EMPC, especially in terms of rejecting the external wind disturbances. The proposed method real-time TCMPC) approach has the advantages of greater robustness and is not heavily dependent upon the accurate dynamics of the model.

Prescribed Performance Active Braking Control with Reference Adaptation for High-Speed Trains

Active braking control systems are vital for the safety of high-speed trains by leading the train operation at its maximum adhesion state. The train adhesion is a nonlinear function of the slip ratio and varies with the uncertain wheel-rail contact conditions. A nonlinear active braking control with rapid and accurate tracking performance is highly required for train braking systems. This paper proposes a novel prescribed performance active braking control with reference adaptation to obtain the maximum adhesion force. The developed feedback linearization controller employs a prescribed performance function that specifies the convergence rate, steady-state error, and maximum overshoot to ensure the transient and steady-state control performance. Furthermore, in the designed control approach, a continuous-time unscented Kalman filter is introduced to estimate the uncertainty of wheel-rail adhesion. The estimation is utilized to represent uncertainty and compensate for the prescribed performance control law. Finally, based on the estimated wheel-rail adhesion, an on-line optimal slip ratio generation algorithm is proposed for the adaptation of the reference wheel slip. The stability of the system is provided, and experiment results validate the effectiveness of the proposed method.

A Biologically Plausible Neural Network for Multichannel Canonical Correlation Analysis.

Cortical pyramidal neurons receive inputs from multiple distinct neural populations and integrate these inputs in separate dendritic compartments. We explore the possibility that cortical microcircuits implement canonical correlation analysis (CCA), an unsupervised learning method that projects the inputs onto a common subspace so as to maximize the correlations between the projections. To this end, we seek a multichannel CCA algorithm that can be implemented in a biologically plausible neural network. For biological plausibility, we require that the network operates in the online setting and its synaptic update rules are local. Starting from a novel CCA objective function, we derive an online optimization algorithm whose optimization steps can be implemented in a single-layer neural network with multicompartmental neurons and local non-Hebbian learning rules. We also derive an extension of our online CCA algorithm with adaptive output rank and output whitening. Interestingly, the extension maps onto a neural network whose neural architecture and synaptic updates resemble neural circuitry and non-Hebbian plasticity observed in the cortex.

Quality management of the chemical-technological process for continuous synthesis of pharmaceutical substances of medicinal compounds in flow microreactors

Objectives. The introduction of digital tools for the development of medicines, intelligent management systems, and quality control is stipulated not only by modern requirements for the chemical and pharmaceutical industry but also by strict regulatory requirements for manufactured products. This principle ensures the release of a quality product on the first attempt. The aim of this study is to develop information support for the intelligent quality management system for the production of active pharmaceutical substances (APSs) for medicines using a fundamentally new technology: continuous synthesis in flow microreactors. To develop the necessary information support, we developed appropriate systemic, informational, and mathematical models; algorithms for the online management of the experiment; and techniques and algorithms to qualitatively assess whether the product meets official regulatory documents.Methods. System analysis techniques, information and mathematical modeling techniques with multireference regression models, and online optimization using the Hook–Jeevs algorithm (a method of expert evaluation based on the concordance factor) were used to solve the problems formulated.Results. To manage the quality of the process of continuous APS synthesis in the flow microreactor, we developed theoretic multiple system models that were designed to build the digital information environment for the process of experimental research. We developed algorithms for mathematical modeling and optimization of the control process based on multiresponse regression models and an online optimization algorithm that allows the process to be managed step by step, taking into account the limitations. Our results show that the degree of conversion is higher in reactions that contain bromodiphenylmethan.Conclusions. Based on mathematical modeling method algorithms for the quality control of the process of continuous APS synthesis on a fundamentally new microreactor system, Qmix were developed. The applicability of the proposed methods and algorithms in the production of the drug diphenhydramine from chlorobenzohydrol and bromobenzohydrol as initial substances was proven by an experimental study. The built models were statistically adequate and valid.

Online Node- and Edge-Deletion Problems with Advice

In online edge- and node-deletion problems the input arrives node by node and an algorithm has to delete nodes or edges in order to keep the input graph in a given graph class Pi at all times. We consider only hereditary properties Pi , for which optimal online algorithms exist and which can be characterized by a set of forbidden subgraphs {{mathcal{F}}} and analyze the advice complexity of getting an optimal solution. We give almost tight bounds on the Delayed Connected{{mathcal{F}}}-Node-Deletion Problem, where all graphs of the family {mathcal{F}} have to be connected and almost tight lower and upper bounds for the DelayedH-Node-Deletion Problem, where there is one forbidden induced subgraph H that may be connected or not. For the DelayedH-Node-Deletion Problem the advice complexity is basically an easy function of the size of the biggest component in H. Additionally, we give tight bounds on the Delayed Connected{mathcal{F}}-Edge-Deletion Problem, where we have an arbitrary number of forbidden connected graphs. For the latter result we present an algorithm that computes the advice complexity directly from {mathcal{F}}. We give a separate analysis for the Delayed ConnectedH-Edge-Deletion Problem, which is less general but admits a bound that is easier to compute.

Observer-based fast nonlinear MPC for multi-DOF maglev positioning system: Theory and experiment

This paper presents an observer-based fast nonlinear model predictive control (NMPC) scheme for translation control of magnetically levitated (maglev) positioning system subject to input saturation. The motivation lies in the improvement of transient characteristics and control performance for positioning systems. The nonlinear dynamical translation model of the maglev positioning system is derived that does not affect the rotation dynamics with special current conditions. The disturbance estimation, obtained by nonlinear disturbance observer, is introduced in the state receding prediction to compensate the errors caused by disturbances and uncertainties. To reduce the computational burden, the stability of the proposed NMPC is established without using any stability-related terminal costs or constraints, and only the short prediction horizon is required for real-time feasibility. The online optimization algorithm underlying the NMPC scheme takes the process constraints into account, and solves the optimal control problem using a parallel structure at each iteration. Comparative experiments are carried out on the positioner to validate the proposed controller has the outperformance in transient/steady-state trajectory tracking, frequency characteristics and robustness against disturbances. The proposed scheme also provides a guidance for the application of NMPC in industrial mechatronic system with fast dynamics.

An online optimization approach to post-disaster road restoration

Natural disasters impact transportation networks adversely and cause road sections to be damaged or blocked. The road network may even become disconnected, impeding accessibility between disaster-stricken areas and critical locations such as hospitals, relief aid depots and transportation hubs. In the immediate response phase, a set of blocked edges should be selected and restored to reconnect the transportation network. While locations of the disrupted roads can be identified using drone or satellite images, an accurate estimation of time to restore a road segment can be carried out only after expert observations on the field. In this article, we study a post-disaster road restoration problem modeled on an undirected edge-weighted graph with k blocked edges, where the unblocking time of a blocked edge is revealed online once the road restoration team visits an end-node of that blocked edge. The objective is to minimize the time at which the road network is reconnected. We first investigate the worst-case performance of online algorithms against offline optimal solutions by means of the competitive ratio. We prove that any online deterministic algorithm cannot achieve a competitive ratio better than 2k−1. We also provide an optimal online algorithm that is proven to achieve this lower bound. In addition, to achieve good performance on realistic instances, we implement an algorithm that solves a mixed integer programming model each time new information is revealed. Since model solution is prohibitively time-consuming, we also propose a novel polynomial time online algorithm. We compare these two algorithms with two other benchmark online algorithms on both Istanbul road network instances and several other city instances from the literature. Our experiments show that the proposed polynomial time online algorithm performs superior to the benchmark ones and obtains solutions close to the offline optimum on all the tested instances.

An Optimal Online Algorithm for Scheduling with Learning Consideration

This paper investigates a classic online scheduling problem with learning effect on a single machine. Specifically, a number of independent jobs that arrive online over time will be processed on a single machine and learning effect implies that the real processing time of job [Formula: see text] is a non-increasing function of its position [Formula: see text], i.e., [Formula: see text], where [Formula: see text] is the basic processing time of job [Formula: see text] and [Formula: see text] is the learning index. Our goal is to minimize the total completion time of all jobs. For the problem, we develop a deterministic polynomial time online algorithm called Delayed Shortest Basic Processing Time (DSBPT) and state that it is an online algorithm with a competitive ratio of 2, which matches the lower bound of the online scheduling problem we focus on.