The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
Stochastic gradient descent (SGD) is one of the most popular numerical algorithms used in machine learning and other domains. Since this is likely to continue for the foreseeable future, it is important to study techniques that can make it run fast on parallel hardware. In this paper, we provide the first analysis of a technique called BUCKWILD! that uses both asynchronous execution and low-precision...
We consider the problem of anomaly localization in a sensor network for multivariate time-series data by computing anomaly scores for each variable separately. To estimate the sparse Gaussian graphical models (GGMs) learned from different sliding windows of the dataset, we propose a new model wherein we constrain sparsity directly through L0 constraint and apply an additional L2 regularization in...
This paper presents a chance-constrained scheduling (CCS) approach for variable wind generation, in the day-ahead timescale, including energy storage. The day-ahead CCS utilizes the ramping of conventional generation as well as the dispatch of energy storage to enhance the load following and ramping support capabilities, to mitigate the impact of net load ramps. The proposed CCS approach is converted...
Restricted Boltzmann Machines (RBMs) have received special attention in the last decade due to their outstanding results in number of applications, such as face and human motion recognition, and collaborative filtering, among others. However, one of the main concerns about RBMs is related to the number of hidden units, which is application-dependent. Infinite RBM (iRBM) was proposed as an alternative...
With the ever-growing popularization of Real Time Bidding (RTB) advertising, the Ad Exchange (AdX) platform has long enjoyed a dominant position in the RTB ecosystem due to its unique role in bridging publishers and advertisers in the supply and demand sides, respectively. A novel technology called header bidding emerged in the recent one or two years, however, is widely believed to have the potential...
In this paper a Markov decision process (MDP) model for virtualized content delivery networks is proposed. We use stochastic optimization to assign cloud site resources to each user group. We propose how quality of experience (QoE) can be included in the modeling and optimization. We then present an optimal solution for a constraint-free version of the problem, and show the improvement in accumulated...
Stochastic optimization is playing an increasingly important in machine learning in the big data era. In this paper, we use forward-backward splitting for the stochastic optimization problems, where the objective is the sum of two functions: one is the expected risk function, another is a regularized term. At each iteration of this method, we just use a single sample to adjust the variables. We prove...
In this paper, the scatter search is modified and combined with the variable-sample technique to deal with the simulation-based optimization problem. First, a new design of scatter search is proposed to deal with the deterministic global optimization problem. Then, the variable-sample technique is combined with the modified scatter search method in order to compose a new global search method that...
A plug-in electric vehicle (PEV) can be used for load shifting household demand using an optimal control strategy to minimize the overall cost of the owner. The PEV can provide initial charge, final desired charge and charging time data to the charging station when plugged in, and the information can be used in a decision-making model to charge/discharge the PEV storage unit in a cost effective manner...
Disturbance rejection is a fundamental problem in control engineering, and there are many methods to achieve a good disturbance rejection. One of the standard methods for the disturbance rejection is to employ an integral compensator. This compensator integrates the error between the reference signal and the output, and use the integrated value with a specific coefficient as a compensating signal...
Simulation Optimization is computationally expensive, especially in large-scale stochastic problem solving, where the computational budget is considered as an important factor. A higher computational budget attempts to generate highly accurate solutions while a lower budget might result in biased or unrealistic solutions. In this paper, the effect of computational budget on the quality of the solution,...
This work investigates a micro immune optimization algorithm originated from the danger theory for single-objective probabilistic constrained optimization without any prior stochastic distribution information. In the whole process of population evolution, the current population is divided into species with different danger levels in terms of constraint dominance and danger radius update. Those species...
The objective of this work is to investigate wind uncertainty within a long term generation and transmission expansion planning framework. Specifically, this work investigates uncertainty associated with wind power within a 2 stage stochastic program and quantifies its importance with the well known quantities “Value of Stochastic Solution” VSS and “Expected Value of Perfect Information” EVPI. Simulations...
This paper deals with a procedure capable to build the scenario tree associated with a parking lot equipped with several bidirectional charging stations for plug-in electric vehicles (EVs). The scenario tree is conceived to be implemented in multistage stochastic optimization models for day-ahead energy management systems of microgrids or, more in general, power distribution networks. Specific operating...
Offering strategy of a price-maker demand response aggregator (DRA) in a two-settlement market is presented in this paper. The aggregator minimizes its cost by offering energy and price bids in the day-ahead market and energy bids in the balancing market. On the other hand, DRA optimally manages the aggregated demands of a large number of electric vehicles and properly distributes them through the...
This paper presents an optimal Day-Ahead Electricity Market (DAM) bidding strategy for an aggregator leveraging a pool of residential prosumers: residential customers with local photovoltaic (PV) production and plug-in electric vehicle (PEV) charging flexibility. The aggregator's point-of-view differs from the social planner angle that is taken in the majority of the existing literature, mainly the...
This paper presents a new framework for risk consideration of participants in electricity markets. Current approaches add risk measures to stochastic problem formulation in order to control the variation of the profits (costs). However, in the proposed method, cumulative distribution function of profit is utilized and it is shaped based on the preferred risk levels of the producer (buyer). The features...
In this paper, a modified stochastic response surface algorithm is proposed for solving expensive black-box global optimization problems. A counter to counts the consecutive failed iterations is used to guide the algorithm to enter into the local search phase or the global search phase accordingly. In the local search phase, the obtained global minimizer of the current response surface model will...
In order to improve the prediction accuracy of GM(1,1) this paper points out the disadvantages of using least square method to solve the parameters of model, attempts to use particle swarm optimization algorithm (PSO) to calculate the parameter of GM(1,1), introduces the stochastic strategy into PSO to endow the inertia weight of particle randomly, and then selects high-rising exponential sequence...
Software-defined cellular networks (SDCN) have been recently introduced to enable flexible cellular network design that facilitates fulfilling 5G design requirements. Placement of controllers within the SDCN plays a crucial role in optimizing its performance. In this paper, we study the controller placement problem in SDCN, considering the uncertainty in cellular user locations. Specifically, our...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.