Reinforcement Learning Based Quantization Strategy Optimal Assignment Algorithm for Mixed Precision

The quantization algorithm compresses the original network by reducing the numerical bit width of the model,which improves the computation speed. Because different layers have different redundancy and sensitivity to databit width. Reducing the data bit width will result in a loss of accuracy. Therefore, it is difficult to determinethe optimal bit width for different parts of the network with guaranteed accuracy. Mixed precision quantizationcan effectively reduce the amount of computation while keeping the model accuracy basically unchanged. In thispaper, a hardware-aware mixed precision quantization strategy optimal assignment algorithm adapted to low bitwidth is proposed, and reinforcement learning is used to automatically predict the mixed precision that meets theconstraints of hardware resources. In the state-space design, the standard deviation of weights is used to measurethe distribution difference of data, the execution speed feedback of simulated neural network accelerator inferenceis used as the environment to limit the action space of the agent, and the accuracy of the quantization model afterretraining is used as the reward function to guide the agent to carry out deep reinforcement learning training. Theexperimental results show that the proposed method obtains a suitable model layer-by-layer quantization strategyunder the condition that the computational resources are satisfied, and themodel accuracy is effectively improved.The proposed method has strong intelligence and certain universality and has strong application potential in thefield of mixed precision quantization and embedded neural network model deployment.

Operation optimization strategy of existing residential building energy-saving renovation market:From the perspective of subject behavior

The energy-saving renovation of existing residential buildings is a crucial measure to achieve the strategic goal of energy conservation and emission reduction in China and build ecologically livable cities.This article focuses on the perspective of subject behavior,starting from analyzing the current situation and difficulties of the operation of the energy-saving renovation market for existing residential buildings in China,drawing on the practical experience of the operation of the existing residential building energy-saving renovation market abroad.Based on principles such as systematicity,humanization,feasibility,and sustainability,the article constructs an operation optimization system of the existing residential building energy-saving renovation market from the perspective of subject behavior.In order to provide a reference for the healthy and orderly operation of the existing residential building energy-saving renovation market,this paper proposes implementation strategies for optimizing the operation of the existing residential building energy-saving renovation market.Suggestions are proposed from four aspects:optimizing the market environment,innovating the financing model,building the information sharing platform,and utilizing the synergies of the main subjects.

Optimization Strategies of Na_(3)V_(2)(PO_(4))_(3) Cathode Materials for Sodium‑Ion Batteries

Na_(3)V_(2)(PO_(4))_(3)(NVP)has garnered great attentions as a prospective cathode material for sodium-ion batteries(SIBs)by virtue of its decent theoretical capacity,superior ion conductivity and high structural stability.However,the inherently poor electronic conductivity and sluggish sodium-ion diffusion kinetics of NVP material give rise to inferior rate performance and unsatisfactory energy density,which strictly confine its further application in SIBs.Thus,it is of significance to boost the sodium storage performance of NVP cathode material.Up to now,many methods have been developed to optimize the electrochemical performance of NVP cathode material.In this review,the latest advances in optimization strategies for improving the electrochemical performance of NVP cathode material are well summarized and discussed,including carbon coating or modification,foreign-ion doping or substitution and nanostructure and morphology design.The foreign-ion doping or substitution is highlighted,involving Na,V,and PO_(4)^(3−)sites,which include single-site doping,multiple-site doping,single-ion doping,multiple-ion doping and so on.Furthermore,the challenges and prospects of high-performance NVP cathode material are also put forward.It is believed that this review can provide a useful reference for designing and developing high-performance NVP cathode material toward the large-scale application in SIBs.

Optimized Strategy for Layout of Crop Production Areas in Hunan Province

The optimized strategy made a comprehensive consideration of resources, technology, market orientation, production scale, industry basis and layout based on the principle of crop security and farmers’ income increasing, and determined the general planning on layout and structure optimization of future crop production ar-eas, with present crop production, market outlook, future industry development, con-cluding crop production characteristics of the 4 crop regions, and proposing function orientation and highlights.

The Networked Manufacturing Resources Optimizing Configuration System and Its Partners Selection Method

The three levels optimizing strategy is put forward for the networked manufacturing resources optimizing configuration, namely, the optimizing of a logical manufacturing process, the optimizing of simulation-based integration of process planning and scheduling, and the optimizing of networked production scheduling. Then, the web services-based architecture of networked manufacturing resources optimizing configuration is brought forward. Finally, the key algorithm of the networked manufacturing resources optimizing configuration is discussed, namely, the two phases manufacturing partners selection method, which including the group technology-based manufacturing resources pre-configuration and the genetic algorithm-based executable manufacturing process optimizing.

Optimal livelihood strategy for different poverty groups among farmers: a case study of the Qin-Ba Mountain area in South-Shaanxi, China

Achieving sustainable livelihood is the ultimate goal of poverty alleviation efforts in mountainous areas,and selecting an optimal livelihood strategy for different poverty-type farmers greatly improves farmers’livelihood capital,resists livelihood risks,and promotes sustainable development.For farmers,optimal livelihood strategy means better employment opportunities,higher family income(or better income structure),and stronger employability or development potential.This paper classifies different types of farmers’poverty on the basis of a quantitative evaluation of farmers’livelihood capital in the Qin-ba Mountain Area in South-Shaanxi by using the k-means clustering method and subsequently the fuzzy evaluation method to evaluate the effectiveness of farmers’livelihood strategies.Then,the multi-attribute decision-making model is used to analyze the selection of optimal livelihood strategies for different poverty-type farmers.The results suggest a significant difference in the selection of the optimal livelihood strategy for different poverty-type farmers.Farmers without financial and human capital choose to"go out to work,"farmers lacking natural capital choose to"acquire social insurance and government relief,"farmers without physical capital choose to"use loans,"and farmers lacking social capital choose to"use savings."Studying the selection of optimal livelihood strategies for different poverty-type farmers can help to propose targeted sustainable livelihood optimization programs for farmers and accelerate efforts to overcome poverty in mountainous areas.

An Optimal Control Strategy for Multi-UAVs Target Tracking and Cooperative Competition

An optimal control strategy of winner-take-all(WTA)model is proposed for target tracking and cooperative competition of multi-UAVs(unmanned aerial vehicles).In this model,firstly,based on the artificial potential field method,the artificial potential field function is improved and the fuzzy control decision is designed to realize the trajectory tracking of dynamic targets.Secondly,according to the finite-time convergence high-order differentiator,a double closed-loop UAV speed tracking the controller is designed to realize the speed control and tracking of the target tracking trajectory.Numerical simulation results show that the designed speed tracking controller has the advantages of fast tracking,high precision,strong stability and avoiding chattering.Finally,a cooperative competition scheme of multiple UAVs based on WTA is designed to find the minimum control energy from multiple UAVs and realize the optimal control strategy.Theoretical analysis and numerical simulation results show that the model has the fast convergence,high control accuracy,strong stability and good robustness.

Data network traffic analysis and optimization strategy of real-time power grid dynamic monitoring system for wide-frequency measurements

The application and development of a wide-area measurement system(WAMS)has enabled many applications and led to several requirements based on dynamic measurement data.Such data are transmitted as big data information flow.To ensure effective transmission of wide-frequency electrical information by the communication protocol of a WAMS,this study performs real-time traffic monitoring and analysis of the data network of a power information system,and establishes corresponding network optimization strategies to solve existing transmission problems.This study utilizes the traffic analysis results obtained using the current real-time dynamic monitoring system to design an optimization strategy,covering the optimization in three progressive levels:the underlying communication protocol,source data,and transmission process.Optimization of the system structure and scheduling optimization of data information are validated to be feasible and practical via tests.

Design of Underwater Robot Lines Based on a Hybrid Automatic Optimization Strategy

In this paper, a hybrid automatic optimization strategy is proposed for the design of underwater robot lines. Isight is introduced as an integration platform. The construction of this platform is based on the user programming and several commercial software including UG6.0, GAMBIT2.4.6 and FLUENT12.0. An intelligent parameter optimization method, the particle swarm optimization, is incorporated into the platform. To verify the strategy proposed, a simulation is conducted on the underwater robot model 5470, which originates from the DTRC SUBOFF project. With the automatic optimization platform, the minimal resistance is taken as the optimization goal;the wet surface area as the constraint condition; the length of the fore-body, maximum body radius and after-body＇s minimum radius as the design variables. With the CFD calculation, the RANS equations and the standard turbulence model are used for direct numerical simulation. By analyses of the simulation results, it is concluded that the platform is of high efficiency and feasibility. Through the platform, a variety of schemes for the design of the lines are generated and the optimal solution is achieved. The combination of the intelligent optimization algorithm and the numerical simulation ensures a global optimal solution and improves the efficiency of the searching solutions.

Optimal control strategy for energy saving in trains under the four-aspect fixed autoblock system

This paper deals with both the leading train and the following train in a train tracking under a four-aspect fixed autoblock system in order to study the optimum operating strategy for energy saving. After analyzing the working principle of the four-aspect fixed autoblock system, an energy-saving control model is created based on the dynamics equation of the Wains. In addition to safety, energy consumption and time error are the main concerns of the model. Based on this model, dynamic speed constraints of the following train are proposed, defined by the leading gain dynamically. At the same time, the static speed constraints defined by the line conditions are also taken into account. The parallel genetic algorithm is used to search the optimum operating strategy. In order to simplify the solving process, the external punishment function is adopted to transform this problem with constraints to the one without constraints. By using the real number coding and the strategy of dividing ramps into three parts, the convergence of GA is accelerated and the length of chromosomes is shortened. The simulation result from a four-aspect fixed autoblock system simulation platform shows that the method can reduce the energy consumption effectively in the premise of ensuring safety and punctuality.

Optimal guidance strategy for flexible load based on hybrid direct load control and time of use

The time-of-use(TOU)strategy can effectively improve the energy consumption mode of customers,reduce the peak-valley difference of load curve,and optimize the allocation of energy resources.This study presents an Optimal guidance mechanism of the flexible load based on strategies of direct load control and time-of-use.First,this study proposes a period partitioning model,which is based on a moving boundary technique with constraint factors,and the Dunn Validity Index(DVI)is used as the objective to solve the period partitioning.Second,a control strategy for the curtailable flexible load is investigated,and a TOU strategy is utilized for further modifying load curve.Third,a price demand response strategy for adjusting transferable load is proposed in this paper.Finally,through the case study analysis of typical daily flexible load curve,the efficiency and correctness of the proposed method and model are validated and proved.

A Novel Two-Level Optimization Strategy for Multi-Debris Active Removal Mission in LEO

Recent studies of the space debris environment in Low Earth Orbit(LEO)have shown that the critical density of space debris has been reached in certain regions.The Active Debris Removal(ADR)mission,to mitigate the space debris density and stabilize the space debris environment,has been considered as a most effective method.In this paper,a novel two-level optimization strategy for multi-debris removal mission in LEO is proposed,which includes the low-level and high-level optimization process.To improve the overall performance of the multi-debris active removal mission and obtain multiple Pareto-optimal solutions,the ADR mission is seen as a Time-Dependant Traveling Salesman Problem(TDTSP)with two objective functions to minimize the total mission duration and the total propellant consumption.The problem includes the sequence optimization to determine the sequence of removal of space debris and the transferring optimization to define the orbital maneuvers.Two optimization models for the two-level optimization strategy are built in solving the multi-debris removal mission,and the optimal Pareto solution is successfully obtained by using the non-dominated sorting genetic algorithm II(NSGA-II).Two test cases are presented,which show that the low level optimization strategy can successfully obtain the optimal sequences and the initial solution of the ADR mission and the high level optimization strategy can efficiently and robustly find the feasible optimal solution for long duration perturbed rendezvous problem.

Strategy Selection for Moving Target Defense in Incomplete Information Game

As a core component of the network,web applications have become one of the preferred targets for attackers because the static configuration of web applications simplifies the exploitation of vulnerabilities by attackers.Although the moving target defense(MTD)has been proposed to increase the attack difficulty for the attackers,there is no solo approach can cope with different attacks;in addition,it is impossible to implement all these approaches simultaneously due to the resource limitation.Thus,the selection of an optimal defense strategy based on MTD has become the focus of research.In general,the confrontation of two players in the security domain is viewed as a stochastic game,and the reward matrices are known to both players.However,in a real security confrontation,this scenario represents an incomplete information game.Each player can only observe the actions performed by the opponent,and the observed actions are not completely accurate.To accurately describe the attacker’s reward function to reach the Nash equilibrium,this work simulated and updated the strategy selection distribution of the attacker by observing and investigating the strategy selection history of the attacker.Next,the possible rewards of the attacker in each confrontation via the observation matrix were corrected.On this basis,the Nash-Q learning algorithm with reward quantification was proposed to select the optimal strategy.Moreover,the performances of the Minimax-Q learning algorithm and Naive-Q learning algorithm were compared and analyzed in the MTD environment.Finally,the experimental results showed that the strategy selection algorithm can enable defenders to select a more reasonable defensive strategy and achieve the maximum possible reward.

Optimal transmission strategy for spatially correlated MIMO systems with channel statistical information

In real multiple-input multiple-output (MIMO) systems, the perfect channel state information (CSI) may be costly or impossible to acquire. But the channel statistical information can be considered relatively stationary during long-term transmission. The statistical information can be obtained at the receiver and fed back to the transmitter and do not require frequent update. By exploiting channel mean and covariance information at the transmitter simultaneously, this paper investigates the optimal trans- mission strategy for spatially correlated MIMO channels. An upper bound of ergodic capacity is derived and taken as the per- formance criterion. Simulation results are also given to show the performance improvement of the optimal transmission strategy.

Control strategy of disc braking systems for downward belt conveyors

Reliability of braking systems is a key requirement to ensure the safety of in using downward belt conveyor brakes. By analyzing and comparing three commonly used braking velocity curves, we conclude that the Harrison curve is the best. Given the characteristics of a downward belt conveyor, we studied the control in a closed-loop velocity, a conventional PID method and an optimal PID control method. We used MATLAB/Simulink to simulate the three control methods. Our simulation results show that opti- mal PID control is especially suitable for disc braking systems. To verif!/the results from theoretical anal- ysis and simulation, a multifunctional test-bed was developed to simulate the braking process of a disc brake system. Our experimental results demonstrate that the optimal PID control can make the output velocity to follow a preset velocity correctly with only small fluctuations, meeting the requirements of a flexible brake for a belt conveyor.

DEVELOPMENT OF THE ENERGY MANAGEMENT STRATEGY FOR PARALLEL HYBRID ELECTRIC URBAN BUSES

A novel parallel hybrid electrical urban bus （PHEUB） configuration consisting of an extra one-way clutch and an automatic mechanical transmission （AMT） is taken as the study subject. An energy management strategy combining a logic threshold approach and an instantaneous optimization algorithm is proposed for the investigated PHEUB. The objective of the energy management strategy is to achieve acceptable vehicle performance and drivability requirements while simultaneously maximizing the engine fuel consumption and maintaining the battery state of charge in its operation range at all times. Under the environment of Matlab/Simulink, a computer simulation model for the PHEUB is constructed by using the model building method combining theoretical analysis and bench test data. Simulation and experiment results for China Typical Bus Driving Schedule at Urban District （CTBDS_UD） are obtained, and the results indicate that the proposed control strategy not only controls the hybrid system efficiently but also improves the fuel economy significantly.

Implementation of Radial Basis Function Artificial Neural Network into an Adaptive Equivalent Consumption Minimization Strategy for Optimized Control of a Hybrid Electric Vehicle

Continued increases in the emission of greenhouse gases by passenger vehicles have accelerated the production of hybrid electric vehicles. With this increase in production, there has been a parallel demand for continuously improving strategies of hybrid electric vehicle control. The goal of an ideal control strategy is to maximize fuel economy while minimizing emissions. Methods exist by which the globally optimal control strategy may be found. However, these methods are not applicable in real-world driving applications since these methods require a priori knowledge of the upcoming drive cycle. Real-time control strategies use the global optimal as a benchmark against which performance can be evaluated. The goal of this work is to use a previously defined strategy that has been shown to closely approximate the global optimal and implement a radial basis function (RBF) artificial neural network (ANN) that dynamically adapts the strategy based on past driving conditions. The strategy used is the Equivalent Consumption Minimization Strategy (ECMS), which uses an equivalence factor to define the control strategy and the power train component torque split. An equivalence factor that is optimal for a single drive cycle can be found offline with a priori knowledge of the drive cycle. The RBF-ANN is used to dynamically update the equivalence factor by examining a past time window of driving characteristics. A total of 30 sets of training data (drive cycles) are used to train the RBF-ANN. For the majority of drive cycles examined, the RBF-ANN implementation is shown to produce fuel economy values that are within ±2.5% of the fuel economy obtained with the optimal equivalence factor. The advantage of the RBF-ANN is that it does not require a priori drive cycle knowledge and is able to be implemented in real-time while meeting or exceeding the performance of the optimal ECMS. Recommendations are made on how the RBF-ANN could be improved to produce better results across a greater array of driving conditions.

The Dragon-shape Strategy of China＇s Regional Economic Development and Policy Analysis

According to this paper, the dragon-shape strategy is the optimized option of China＇s future strategy with respect to the geographic distribution of regional economy.

Maneuver strategy of lifting reentry vehicle

An optimal maneuver strategy is proposed for lifting reentry vehicle to reach the maximum lateral range after reentering the atmosphere. Aiming at problems that too many co-state variables and difficulty in estimating the initial values of co-state variables,the equilibrium glide condition (EGC) is utilized to reduce the reentry motion equations and then the optimal maneuver strategy satisfied above performance index is derived. This maneuvering strategy is applied to the lifting reentry weapon platform CAV which was designed by America recently to realize both longitudinal and lateral trajectory design by controlling the attack angle and the bank angle respectively. The simulation result indicates that the maneuver strategy proposed enables CAV to reach favorable longitudinal range and lateral range.

OPTIMAL FEED STRATEGY FOR FED-BATCH GLYCEROL FERMENTATION DETERMINED BY MAXIMUM PRINCIPLE

Optimal glucose feed strategy for glycerol fed-batch fermentation was investigated by Pontryagin’s maximum principle to maximize the final glycerol yield. The problem was solved by a nonsingular control approach by selecting the culture volume as the control variable, then the general optimal feed profile was numerically determined.