检索结果-内蒙古大学图书馆

Heterogeneous multi-agent resource allocation through multi-bidding with applications to precision agriculture⁎

IFAC-PapersOnLine 2018年第23期51卷 194-199页

作者： Cobbenhagen, A.T.J.R. Antunes, D.J. van de Molengraft, M.J.G. Heemels, W.P.M.H. Control Systems Technology group Dept. of Mechanical Engineering Eindhoven University of Technology Eindhoven Netherlands

In this paper we consider the problem of allocating multiple resources to a number of clients by a group of heterogeneous agents over time such that the clients can produce products while maximizing a profit function. We propose an approximate optimization framework in which every client provides multiple bids from which the agents choose such that an allocation is feasible and that the profit function is maximized over time. The proposed framework exploits decomposition techniques that can be used for large-scale multi-agent resource allocation problems in which the cost objective is additive, the dynamics of product generation is non-linear and the agents have different capabilities. Interestingly, the decomposition can be solved in a distributed fashion, enabling application to large-scale problems. We apply this decomposition to the management of resources and agents in precision agriculture as an inspirational and important application domain of the obtained results. We show that our framework can be used in order to schedule the time, location and quantity of resources that every agent must provide whilst optimizing the profit of the entire farm over the growing season. © 2018

关键词： dynamic programming Agriculture Multi agent systems Precision agriculture Profitability Resource allocation approximate dynamic programming approximate optimization Decomposition technique Distributed optimization Heterogeneous agents Multi agent Optimal controls Resource allocation problem

来源：评论

学校读者我要写书评

暂无评论

Management of a stage-structured insect pest: an application of approximate optimization

引用

ECOLOGICAL APPLICATIONS 2018年第4期28卷 938-952页

作者： Hackett, Sean C. Bonsall, Michael B. Univ Oxford Dept Zool Math Ecol Res Grp South Parks Rd Oxford OX1 3PS England St Peters Coll New Inn Hall St Oxford OX1 2DL England

Ecological decision problems frequently require the optimization of a sequence of actions over time where actions may have both immediate and downstream effects. dynamic programming can solve such problems only if the dimensionality is sufficiently low. approximate dynamic programming (ADP) provides a suite of methods applicable to problems of arbitrary complexity at the expense of guaranteed optimality. The most easily generalized method is the look-ahead policy: a brute-force algorithm that identifies reasonable actions by constructing and solving a series of temporally truncated approximations of the full problem over a defined planning horizon. We develop and apply this approach to a pest management problem inspired by the Mediterranean fruit fly, Ceratitis capitata. The model aims to minimize the cumulative costs of management actions and medfly-induced losses over a single 16-week season. The medfly population is stage-structured and grows continuously while management decisions are made at discrete, weekly intervals. For each week, the model chooses between inaction, insecticide application, or one of six sterile insect release ratios. Look-ahead policy performance is evaluated over a range of planning horizons, two levels of crop susceptibility to medfly and three levels of pesticide persistence. In all cases, the actions proposed by the look-ahead policy are contrasted to those of a myopic policy that minimizes costs over only the current week. We find that look-ahead policies always out-performed a myopic policy and decision quality is sensitive to the temporal distribution of costs relative to the planning horizon: it is beneficial to extend the planning horizon when it excludes pertinent costs. However, longer planning horizons may reduce decision quality when major costs are resolved imminently. ADP methods such as the look-ahead-policy-based approach developed here render questions intractable to dynamic programming amenable to inference but should be

关键词： approximate dynamic programming Ceratitis capitata continuous state space dynamic programming look-ahead policy sequential decision problem stage structure sterile insect release

来源：评论

学校读者我要写书评

暂无评论

Learning Output Reference Model Tracking for Higher-Order Nonlinear Systems with Unknown dynamics

引用

ALGORITHMS 2019年第6期12卷 121-121页

作者： Radac, Mircea-Bogdan Lala, Timotei Politehn Univ Timisoara Dept Automat & Appl Informat 2 Bd V Parvan Timisoara 300223 Romania

Linearly and nonlinearly parameterized approximate dynamic programming approaches used for output reference model (ORM) tracking control are proposed. The ORM tracking problem is of significant interest in practice since, with a linear ORM, the closed-loop control system is indirectly feedback linearized and value iteration (VI) offers the means to achieve ORM tracking without using process dynamics. Ranging from linear to nonlinear parameterizations, a successful approximate VI implementation for continuous state-action spaces depends on several key parameters such as: problem dimension, exploration of the state-action space, the state-transitions dataset size, and suitable selection of the function approximators. We show that using the same transitions dataset and under a general linear parameterization of the Q-function, high performance ORM tracking can be achieved with an approximate VI scheme, on the same performance level as that of a neural-network (NN)-based implementation that is more complex and takes significantly more time to learn. However, the latter proves to be more robust to hyperparameters selection, dataset size, and to exploration strategies, recommending it as the de facto practical implementation. The case study is aimed at ORM tracking of a real-world nonlinear two inputs-two outputs aerodynamic process with ten internal states, as a representative high order system.

关键词： approximate dynamic programming reinforcement learning data-driven control model-free control reference trajectory tracking output reference model multivariable control aerodynamic rotor system neural networks learning systems

来源：评论

学校读者我要写书评

暂无评论

A Bayesian Machine Learning Approach for Optimizing dynamic Treatment Regimes

引用

JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION 2018年第523期113卷 1255-1267页

作者： Murray, Thomas A. Yuan, Ying Thall, Peter F. Univ Texas MD Anderson Canc Ctr Dept Biostat Houston TX 77030 USA

Medical therapy often consists of multiple stages, with a treatment chosen by the physician at each stage based on the patient's history of treatments and clinical outcomes. These decisions can be formalized as a dynamic treatment regime. This article describes a new approach for optimizing dynamic treatment regimes, which bridges the gap between Bayesian inference and existing approaches, like Q-learning. The proposed approach fits a series of Bayesian regression models, one for each stage, in reverse sequential order. Each model uses as a response variable the remaining payoff assuming optimal actions are taken at subsequent stages, and as covariates the current history and relevant actions at that stage. The key difficulty is that the optimal decision rules at subsequent stages are unknown, and even if these decision rules were known the relevant response variables may be counterfactual. However, posterior distributions can be derived from the previously fitted regression models for the optimal decision rules and the counterfactual response variables under a particular set of rules. The proposed approach averages over these posterior distributions when fitting each regression model. An efficient sampling algorithm for estimation is presented, along with simulation studies that compare the proposed approach with Q-learning. Supplementary materials for this article are available online.

关键词： approximate dynamic programming Backward induction Bayesian additive regression trees Gibbs sampling Potential outcomes

来源：评论

学校读者我要写书评

暂无评论

Path-based dynamic Pricing for Vehicle Allocation in Ridesharing Systems with Fully Compliant Drivers

引用

Transportation Research Procedia 2019年 38卷 77-97页

作者： Chao Lei Zhoutong Jiang Yanfeng Ouyang Department of Civil and Environmental Engineering University of Illinois at Urbana-Champaign Urbana IL 61801 USA

Rapidly advancing on-demand ridesharing services, including those with self-driving technologies, hold the promise to revolutionize delivery of mobility. Yet, significant imbalance between spatiotemporal distributions of vehicle supply and travel demand poses a pressing challenge. This paper proposes a multi-period game-theoretic model that addresses dynamic pricing and idling vehicle dispatching problems in the on-demand ridesharing systems with fully compliant drivers/vehicles. A dynamic mathematical program with equilibrium constraints (MPEC) is formulated to capture the interdependent decision-making processes of the mobility service provider (e.g., regarding vehicle allocation) and travelers (e.g., regarding ride-sharing and travel path options). An algorithm based on approximate dynamic programming (ADP), with customized subroutines for solving the MPEC, is developed to solve the overall problem. It is shown with numerical experiments that the proposed dynamic pricing and vehicle dispatching strategy can help ridesharing service providers achieve better system performance (as compared with myopic policies) while facing spatial and temporal variations in ridesharing demand.

关键词： Ridesharing dynamic pricing Self-driving vehicle Bi-level optimization MPEC approximate dynamic programming

来源：评论

学校读者我要写书评

暂无评论

Shape Constraints in Economics and Operations Research

引用

STATISTICAL SCIENCE 2018年第4期33卷 527-546页

作者： Johnson, Andrew L. Jiang, Daniel R. Texas A&M Univ Dept Ind & Syst Engn College Stn TX 77840 USA Osaka Univ Sch Informat Sci & Technol Suita Osaka 5650871 Japan Univ Pittsburgh Dept Ind Engn Pittsburgh PA 15261 USA

Shape constraints, motivated by either application-specific assumptions or existing theory, can be imposed during model estimation to restrict the feasible region of the parameters. Although such restrictions may not provide any benefits in an asymptotic analysis, they often improve finite sample performance of statistical estimators and the computational efficiency of finding near-optimal control policies. This paper briefly reviews an illustrative set of research utilizing shape constraints in the economics and operations research literature. We highlight the methodological innovations and applications, with a particular emphasis on utility functions, production economics and sequential decision making applications.

关键词： Shape constraints multivariate convex regression nonparametric regression production economics consumer preferences revealed preferences approximate dynamic programming reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Automated Feature Synthesis From Relational Database For Data Science Related Problems 16

Automated Feature Synthesis From Relational Database For Dat...

引用

16th International Conference on Frontiers of Information Technology (FIT)

作者： Fatima, Afifa Khan, Faizan Ali Raza, Aneeq Kamran, Abdul Basit Sir Syed CASE Inst Technol Comp Sci Islamabad Pakistan

ISBN: (纸本)9781538693551

Feature Synthesis is the most important aspect of Data Science. This process requires domain knowledge of the data to produce meaningful features that support decision-making algorithms. The biggest technical obstacle which a data scientist must subdue is the synthesis of such features that would assist AI algorithms to predict the target variable with the least error. Feature synthesis is a crucial step toward the application of machine learning which consumes time and resources. Automated Feature Synthesis obviates the tedious task of data scientists. Data Science Machine and One Button Machine are pioneering researches that explored the possibility of feature synthesis from raw relational data by applying linear exploratory algorithms which increase the dimensionality of the feature matrix exponentially, and hence, they consume a lot of time and memory. Our research revolves around the process of feature synthesis from the structured and unstructured data types for relational data which will provide an effective automation of a hectic manual task for the non-experts consuming a reasonable amount of resources. We devised a solution to reduce the time complexity of the operation by reducing dimensionality while considering the target variable by using approximate dynamic programming. The model is designed utilizing the concept of sparse reward setting, and it tries to achieve maximum reward as it explores the feature set. The target variable plays its part to compute the reward eventually leading to an optimum feature set. This process reduces data exploration time, cost and effort for both data scientists and non-experts.

关键词： approximate dynamic programming Automated Feature Synthesis

来源：评论

学校读者我要写书评

暂无评论

ADP-based Home Energy Management System: A Case Study using DYNAMO

ADP-based Home Energy Management System: A Case Study using ...

引用

IEEE-Power-and-Energy-Society General Meeting (PESGM)

作者： Faqiry, M. N. Wang, L. Wu, H. Krishnamurthy, D. Palmintier, B. Kansas State Univ Dept Elect & Comp Engn Manhattan KS 66506 USA Natl Renewable Energy Lab Power Syst Engn Ctr Golden CO 80401 USA

ISBN: (纸本)9781538677032

This paper explores Home Energy Management System (HEMS) algorithms to minimize household cost while maintaining comfort when faced with uncertain weather, and demand. Specifically, we consider a HEMS that optimizes forward looking schedules for a home's heating, ventilation, and air conditioning (HVAC);water heater (WH);and electric vehicle (EV) charging while considering uncertainty in outside temperature, hot water usage, and non-controllable load (NCL). We adopt a dynamic programming (DP) formulation and utilize the dynamic programming for Adaptive Modeling and Optimization (DYNAMO) toolkit to implement DP and approximate dynamic programming (ADP) algorithms. Simulation results under a single tariff plan compare the quality of the solution generated by ADP to that of DP, and show significant improvement in computation time while maintaining acceptable solution accuracy.

关键词： approximate dynamic programming demand response electric vehicle home energy management system stochastic mixed-integer programming

来源：评论

学校读者我要写书评

暂无评论

The Military Inventory Routing Problem: Utilizing Heuristics Within a Least Squares Temporal Differences Algorithm to Solve a Multiclass Stochastic Inventory Routing Problem with Vehicle Loss

The Military Inventory Routing Problem: Utilizing Heuristics...

引用

作者： Ethan L. Salgado AIR FORCE INSTITUTE OF TECHNOLOGY

学位级别：博士

Military commanders currently resupply forward operating bases (FOBs) from a central location within an area of operations mainly via convoy operations in a way that closely resembles vendor managed inventory practices. Commanders must decide when and how much inventory to distribute throughout their area of operations while minimizing soldier risk. Technology currently exists that makes utilizing unmanned cargo aerial vehicles (CUAVs) for resupply an attractive alternative due to the dan– gers of utilizing convoy operations. Enemy actions in wartime environments pose a significant risk to a CUAV's ability to safely deliver supplies to a FOB. This dis– sertation develops a Markov decision process (MDP) model to examine this military inventory routing problem (MILIRP). The first paper examines the structure of the MILIRP by considering a small problem instance and prove value function monotonicity when a sufficient penalty is applied. Moreover, this paper develops a monotone least squares temporal differ– ences (MLSTD) algorithm that exploits this structure and demonstrate its efficacy for approximately solving this problem class. This work compares MLSTD to least squares temporal differences (LSTD), a similar ADP algorithm that does not ex– ploit monotonicity. MLSTD attains a 3.05% optimality gap for a baseline scenario and outperforms LSTD by 31.86% on average in our computational experiments. The second paper expands the problem complexity with additional FOBs. This work gen– erates two new algorithms, Index and Rollout, for the routing portion and implement an LSTD algorithm that utilized these to produce solutions 22% better than myopic generated solutions on average. The third paper greatly increases problem complex– ity with the addition of supply classes. This research formulates an MDP model to handle the increased complexity and implement our LSTD-Index and LSTD-Rollout algorithms to solve this larger problem instance and perform 21% better on average than

关键词： stochastic inventory routing approximate dynamic programming Markov decision process vendor managed inventory least squares temporal differences

来源：评论

学校读者我要写书评

暂无评论

Stochastic dynamic programming approach to managing power system uncertainty with distributed storage

引用

COMPUTATIONAL MANAGEMENT SCIENCE 2018年第1期15卷 87-110页

作者： Zephyr, Luckny Anderson, C. Lindsay Cornell Univ Dept Biol & Environm Engn Riley Robb Hall Ithaca NY 14853 USA

Wind integration in power grids is challenging because of the uncertain nature of wind speed. Forecasting errors may have costly consequences. Indeed, power might be purchased at highest prices to meet the load, and in case of surplus, power may be wasted. Energy storage may provide some recourse against the uncertainty of wind generation. Because of their sequential nature, in theory, power scheduling problems may be solved via stochastic dynamic programming. However, this scheme is limited to small networks by the so-called curse of dimensionality. This paper analyzes the management of a network composed of conventional power units and wind turbines through approximate dynamic programming, more precisely stochastic dual dynamic programming. A general power network model with ramping constraints on the conventional generators is considered. The approximate method is tested on several networks of different sizes. The numerical experiments also include comparisons with classical dynamic programming on a small network. The results show that the combination of approximation techniques enables to solve the problem in reasonable time.

关键词： Power grid management Energy storage Stochastic dynamic programming Stochastic dual dynamic programming approximate dynamic programming Generalized linear programming

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：