Method for regulating and controlling visual comfort of intelligent sunshade shutter based on dynamic programming algorithm

Abstract

The invention provides an intelligent sunshade shutter visual comfort performance regulation and control method based on a dynamic programming algorithm. The method comprises the following steps: step 1, collecting potential influence factors of visual comfort performance of a light environment and visual comfort evaluation index values, and constructing a visual comfort data set of the light environment of the office space; step 2, training a random forest model algorithm in machine learning by using the data set in the step 1 to complete construction of a light environment visual comfort index prediction model; and step 3, constructing a sunshade regulation and control intelligent body based on dynamic programming by using the light environment visual comfort index prediction model in the step 2, and completing intelligent sunshade shutter visual comfort performance regulation and control based on a dynamic programming algorithm. According to the method, various factors of the light environment and the visual comfort evaluation indexes are comprehensively considered, intelligent regulation and control of the sunshade shutter are achieved by means of a dynamic planning algorithm, the visual comfort requirements of people in different light environments can be met to the maximum extent, and the visual comfort of the indoor light environment is practically and effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of building intelligence and light environment control technology, and in particular to a method for controlling the visual comfort performance of intelligent sunshade blinds based on a dynamic programming algorithm. The method is widely applicable to office building spaces, and its purpose is to optimize the indoor light environment by intelligently controlling the sunshade blinds, thereby improving people's visual comfort in the indoor environment. Background Art

[0002] In modern building systems, indoor natural lighting environment has a critical impact on people's visual perception, work efficiency, and physical and mental health. Appropriate natural lighting conditions can effectively reduce visual fatigue and improve work and study efficiency; conversely, a poor light environment may cause a series of problems such as eye fatigue and headaches. As a common building shading facility, sunshade blinds play an important role in regulating the amount of light entering the room and the light distribution, and are a key factor in improving the indoor light environment.

[0003] Traditional sunshade shutter control methods mainly rely on manual operation or simple timing control, and cannot be dynamically adjusted according to real-time changes in the light environment and people's visual comfort needs. Although some existing intelligent sunshade systems can be controlled based on a single parameter such as light intensity, they lack comprehensive consideration of multiple factors of the light environment and visual comfort evaluation indicators, making it difficult to truly optimize visual comfort performance.

[0004] Dynamic programming algorithm is a mathematical method for solving optimization problems in multi-stage decision-making processes, and is widely used in many fields such as resource allocation and path planning. However, there is currently no technology that combines dynamic programming algorithm with the light environment visual comfort index prediction model and applies it to intelligent sunshade blinds control. Summary of the invention

[0005] The purpose of the present invention is to solve the problems in the prior art and propose a method for controlling the visual comfort performance of intelligent sunshade blinds based on a dynamic programming algorithm. The method can comprehensively consider various factors of the light environment and visual comfort evaluation indicators, and realize intelligent control of the sunshade blinds through a dynamic programming algorithm, thereby maximizing the visual comfort of the indoor light environment.

[0006] The present invention is implemented by the following technical scheme. The present invention proposes a method for controlling the visual comfort performance of intelligent sunshade blinds based on a dynamic programming algorithm. The method comprises the following steps:

[0007] Step 1: In the selected typical office building space, using the data simulation method, collect the potential influencing factors of the light environment visual comfort performance and the visual comfort evaluation index values, and construct the office space light environment visual comfort data set;

[0008] Step 2: According to the non - linear correlation between the potential influencing factors of visual comfort performance in the light environment and the values of visual comfort evaluation indicators, use the dataset in Step 1 to train the random forest model algorithm in machine learning to complete the construction of the visual comfort index prediction model for the light environment; the process of model construction includes pre - processing of the dataset, splitting of the test set and the training set, and model training and verification.

[0009] Step 3: Take the visual comfort evaluation index as the optimization goal, use the visual comfort index prediction model for the light environment in Step 2, and combine with the dynamic programming algorithm to develop the visual comfort performance regulation of intelligent sunshade louvers based on the dynamic programming algorithm; specifically, take the visual comfort index prediction model for the light environment in Step 2 as the state model, obtain the state transition function and the reward function, and construct a sunshade regulation intelligent agent based on dynamic programming.

[0010] Furthermore, in the process of data collection and model construction,

[0011] First, collect the data of the potential influencing factors of visual comfort performance in the light environment and the corresponding values of visual comfort evaluation indicators, and construct a dataset based on this.

[0012] Second, carry out pre - processing work on the dataset, including data cleaning, encoding, and standardization.

[0013] Finally, divide the pre - processed dataset into a training set and a test set according to a certain proportion, use the random forest model algorithm to train the training set, construct a visual comfort index prediction model for the light environment, and verify the accuracy of the model through the test set.

[0014] Furthermore, the problem modeling of the regulation method is specifically as follows:

[0015] Define the state space: The state includes various parameters of the light environment, the current angle of the sunshade louver, and time information. Use s to represent the state, and the state space S is the set of all possible states.

[0016] Define the action space: The action refers to adjusting the angle of the sunshade louver. Use a to represent the action, and the action space A is the set of all possible actions.

[0017] Determine the optimization goal: The optimization goal is set to maximize the visual comfort evaluation index, and the output of the visual comfort index prediction model for the light environment is used as the value function V(s).

[0018] Furthermore, the state transition function: Regard the visual comfort index prediction model for the light environment as the state model. Given the current state s and action a, obtain the predicted value of the next state s' through this model; if the state transition is deterministic, P(s'|s,a) = 1 if and only if s' is the next state predicted by the model, otherwise P(s'|s,a) = 0.

[0019] Further, the reward function: The reward function R(s, a) is defined as the change value of the visual comfort evaluation index of state s after taking action a, that is, R(s, a) = V(s') - V(s), where s' is the next state transferred to after taking action a.

[0020] Further, during the process of constructing a sunshade control agent based on dynamic programming, initialization is performed. The value function is initialized to zero and applies to all states; the value function represents the expected value of the long-term cumulative reward that can be obtained by following the optimal policy starting from the initial state; the initial policy is a random policy, and a random action is selected for each state. Specifically, the value function V(s) is initialized to zero for all states s ∈ S; the initial policy π(s) is initialized to a random policy, that is, for each state s, a random action a ∈ A is selected.

[0021] Further, during the process of constructing a sunshade control agent based on dynamic programming, the dynamic programming algorithm is iterated: The value function is calculated according to the current policy, and the value function is updated for each state using the Bellman equation; when improving the policy, the policy is improved according to the current value function; for each state, the action that maximizes the action value function is selected as the new policy; the policy evaluation and policy improvement steps are repeated until the policy converges, that is, the policy no longer changes.

[0022] Further, the specific process of intelligent sunshade control implementation is as follows:

[0023] Real-time state monitoring: Sensors are used to monitor the parameters of the light environment and the current state of the sunshade louvers in real time, and this information is input as the current state s into the sunshade control agent based on dynamic programming.

[0024] Action decision-making: Based on the current state s, the agent selects an action a through the converged optimal policy π(s), that is, determines the best adjustment angle of the sunshade louvers.

[0025] Action execution: The selected action a is sent to the sunshade louver control system to perform the angle adjustment operation of the sunshade louvers.

[0026] Loop iteration: Repeat the real-time state monitoring, action decision-making, and action execution steps to achieve real-time dynamic control of the sunshade louvers.

[0027] The present invention also proposes an electronic device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the intelligent sunshade louver visual comfort performance control method based on the dynamic programming algorithm are implemented.

[0028] The present invention also provides a computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the steps of the intelligent shading louver visual comfort performance regulation method based on the dynamic programming algorithm.

[0029] Advantages of the present invention:

[0030] 1. Improve visual comfort: The present invention comprehensively considers various factors of the light environment and visual comfort evaluation indicators, and realizes intelligent regulation of the shading louvers by means of the dynamic programming algorithm, which can best meet people's visual comfort needs in different light environments and effectively improve the visual comfort of the indoor light environment.

[0031] 2. Adaptive dynamic adjustment: This method can dynamically adjust the angle of the shading louvers according to the real-time changes in the light environment and the state of the shading louvers, realizing adaptive light environment regulation, and significantly improving the accuracy and flexibility of the regulation.

[0032] 3. Energy-saving and efficient: By reasonably controlling the angle of the shading louvers, unnecessary direct sunlight is reduced, and the energy consumption of indoor air conditioners and lighting equipment is lowered, achieving the dual goals of energy conservation and comfort. Description of the drawings

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0034] Figure 1 It is a flowchart of an intelligent shading louver visual comfort performance regulation method based on the dynamic programming algorithm of the present invention.

[0035] Figure 2 It is a regulation process diagram of an intelligent shading louver visual comfort performance regulation method based on the dynamic programming algorithm. Detailed implementation manners

[0036] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0037] Combined with Figure 1 - Figure 2, the present invention proposes an intelligent shading louver visual comfort performance regulation method based on the dynamic programming algorithm, and the method includes the following steps:

[0038] Step 1, in the selected typical office building space, use the data simulation method to collect the potential influencing factors of the light environment visual comfort performance and the visual comfort evaluation index values, and construct a light environment visual comfort dataset for the office space;

[0039] In Step 1, the data simulation method selects

[0040] Tools: Use professional lighting simulation software (such as Radiance, DIVA, or EnergyPlus) to generate light environment data, and combine with parametric modeling tools (such as Grasshopper) to achieve automated simulation.

[0041] Fixed parameter settings: Cover variables such as sky model type, Radiance simulation parameters, building geometric dimensions, materials, orientations, window sizes, and space layouts.

[0042] Verification: Collect real data through on-site sensors (illuminance meters, glare sensors), calibrate the simulation results, and ensure that the error is within 10%.

[0043] In Step 1, the influencing factors and evaluation indicators

[0044] Input features: Season, date, all-day time, climate parameters (outdoor horizontal normal direct irradiance, outdoor diffuse irradiance, dry bulb temperature), and the angle of the shading louver blades.

[0045] Output index: sUDI, that is, the ratio of the measurement points that meet the effective horizontal illuminance to all measurement points at a certain moment, or the area ratio that meets the effective horizontal illuminance.

[0046] In Step 1, the sUDI simulation based on the simulation platform

[0047] Run the simulation: Traverse the representative dates and time periods of the whole year to start the simulation, which can be the representative time periods in typical summer and winter.

[0048] Construct the dataset: Simulate and generate UDI data, form a mapping with the feature data, and construct a dataset for machine learning training.

[0049] Step 2, according to the non-linear association between the potential influencing factors of the light environment visual comfort performance and the visual comfort evaluation index values, use the dataset in Step 1 to train the random forest model algorithm in machine learning to complete the construction of the light environment visual comfort index prediction model; the process of model construction includes preprocessing of the dataset, splitting of the test set and training set, and model training and verification;

[0050] In Step 2, data preprocessing

[0051] Data encoding: Convert it into a numerical type for the machine learning model to process. Specifically, label encoding is required for seasonal and date data.

[0052] Data standardization: Standardize the numerical features so that different features have the same scale.

[0053] In Step 2, splitting the test set and the training set

[0054] Determine the splitting ratio: Usually, the dataset is divided into a training set and a test set according to a certain ratio. Common ratios are 7:3 or 8:2. The training set is used for model training, and the test set is used to evaluate the generalization ability of the model.

[0055] Random splitting: Use a random function algorithm to randomly permute the samples and split the test set and the training set according to the set splitting ratio.

[0056] In Step 2, model training and validation

[0057] Model initialization: For sUDI, call the random forest regression algorithm, initialize the random forest model, and set the corresponding hyperparameters.

[0058] Model training: Use the training set data to train the random forest model, passing in the feature matrix and the target variable of the training set.

[0059] Model validation: Use the trained model to predict the test set, call the prediction method of the model, and pass in the feature matrix of the test set.

[0060] Calculation of evaluation metrics: Evaluate the random forest regression model. Commonly used evaluation metrics include mean squared error (MSE), root mean squared error (RMSE), mean absolute error (MAE), and coefficient of determination.

[0061] In Step 2, model optimization and tuning

[0062] Hyperparameter optimization: Use methods such as Grid Search or Random Search to tune the hyperparameters of the random forest model to find the optimal parameter combination and improve the performance of the model.

[0063] Model training with optimal parameters: After obtaining the optimal parameters, use the optimal parameters to train the model.

[0064] Step 3: Taking the visual comfort evaluation index as the optimization objective, using the visual comfort index prediction model of the light environment in Step 2, and combining with the dynamic programming algorithm to develop the visual comfort performance regulation of intelligent sunshade louvers based on the dynamic programming algorithm. Specifically, taking the visual comfort index prediction model of the light environment in Step 2 as the state model, obtaining the state transition function and the reward function, and constructing a sunshade regulation intelligent agent based on dynamic programming.

[0065] In Step 3, modeling the visual comfort problem of the light environment

[0066] Define the state space: The state includes season, date, time of day, climate parameters (outdoor horizontal normal direct irradiance, outdoor diffuse irradiance, dry bulb temperature), and the angle of the sunshade louver blades.

[0067] Define the action space: An action is an operation that the intelligent agent can take. For the sunshade louver, the action is to adjust the angle of the louver.

[0068] Determine the optimization function: The optimization objective is to maximize the visual comfort evaluation index. Take the sUDI of the visual comfort index prediction model of the light environment obtained in Step 2 as the optimization function.

[0069] In Step 3, obtain the state transition function and the reward function

[0070] State transition function: The state transition function describes the probability that the system transfers to the next state s' after taking action a in the current state s. Based on the visual comfort index prediction model of the light environment in Step 2, use it as the state model to obtain the state transition function.

[0071] Since the visual comfort index prediction model of the light environment in Step 2 can predict the future light environment based on the current light environment parameters and the state of the sunshade louver, this model can be used as the state model to obtain the reward function.

[0072] In Step 3, construct a sunshade regulation intelligent agent based on dynamic programming

[0073] Initialization: Initialize the value function to zero for all states. The value function represents the expected value of the long-term cumulative reward that can be obtained by following the optimal policy starting from the initial state; while the initial policy is a random policy, randomly select an action for each state.

[0074] Dynamic programming algorithm iteration: Calculate the value function according to the current policy. Update the value function for each state using the Bellman equation. When improving the policy, improve the policy according to the current value function. For each state, select the action that maximizes the action value function as the new policy. Repeat the policy evaluation and policy improvement steps until the policy converges, that is, the policy no longer changes.

[0075] A method for regulating the visual comfort performance of intelligent shading blinds based on the dynamic programming algorithm realizes the process of regulating the visual comfort performance of intelligent shading blinds based on the dynamic programming algorithm. The specific process is as follows:

[0076] 1. Data collection and model construction

[0077] 1.1 Widely collect data on the potential influencing factors of the visual comfort performance of the light environment and the corresponding visual comfort evaluation index values, and construct a data set based on this.

[0078] 1.2 Carry out preprocessing work on the data set, covering operation processes such as data cleaning, encoding, and standardization.

[0079] 1.3 Divide the preprocessed data set into a training set and a test set according to a certain proportion, use the random forest model algorithm to train the training set, construct a prediction model for the visual comfort index of the light environment, and verify the accuracy of the model through the test set.

[0080] 2. Problem modeling

[0081] 2.1 Define the state space: The state includes various parameters of the light environment, the current angle of the shading blinds, time, etc. Use s to represent the state, and the state space S is the set of all possible states.

[0082] 2.2 Define the action space: The action refers to adjusting the angle of the shading blinds. Use a to represent the action, and the action space A is the set of all possible actions.

[0083] 2.3 Determine the optimization goal: The optimization goal is set to maximize the visual comfort evaluation index, and the output of the prediction model for the visual comfort index of the light environment is used as the value function V(s).

[0084] 3. Obtain the state transition function and the reward function

[0085] 3.1 State transition function: Regard the prediction model for the visual comfort index of the light environment as the state model. Given the current state s and the action a, obtain the predicted value of the next state s' through this model. If the state transition is deterministic, P(s'|s,a) = 1 if and only if s' is the next state predicted by the model, otherwise P(s'|s,a) = 0.

[0086] 3.2 Reward function: Define the reward function R(s,a) as the change value of the visual comfort evaluation index of the state s after taking the action a, that is, R(s,a) = V(s') - V(s), where s' is the next state transferred to after taking the action a.

[0087] 4. Construct a shading control intelligent agent based on dynamic programming

[0088] 4.1 Initialization: Initialize the value function V(s) to zero for all states s ∈ S; initialize the policy π(s) to a random policy, i.e., for each state s, randomly select an action a ∈ A.

[0089] 4.2 Iterate the dynamic programming algorithm, including policy evaluation and policy improvement.

[0090] 4.2.1 Policy evaluation: Calculate the value function V(s) according to the current policy π(s), and update the value function using the Bellman equation: V(s) = Σ a∈A π(a|s)∑ s'∈S P(s'|s,a)[R(s,a) + γV(s')], where γ is the discount factor, usually taking values in the range [0, 1].

[0091] 4.2.2 Policy improvement: Improve the policy π(s) according to the current value function V(s). For each state s, select the action a that maximizes the action value function Q(s,a) as the new policy, i.e.,

[0092]

[0093] 4.2.3 Repeatedly perform the policy evaluation and policy improvement steps until the policy converges, that is, the policy no longer changes.

[0094] 5. Intelligent sunshade regulation implementation

[0095] 5.1 Real-time state monitoring: Use sensors to monitor the parameters of the light environment and the current state of the sunshade louvers in real time, and input this information as the current state s into the sunshade regulation intelligent agent based on dynamic programming.

[0096] 5.2 Action decision-making: Based on the current state s, the intelligent agent selects an action a through the converged optimal policy π(s), that is, determines the best adjustment angle of the sunshade louvers.

[0097] 5.3 Action execution: Send the selected action a to the sunshade louver control system to execute the angle adjustment operation of the sunshade louvers.

[0098] 5.4 Loop iteration: Repeat the real-time state monitoring, action decision-making, and action execution steps to achieve real-time dynamic regulation of the sunshade louvers.

[0099] Embodiment

[0100] The present invention proposes an intelligent sunshade louver visual comfort performance regulation method based on a dynamic programming algorithm, and the method includes:

[0101] Step 1: Data collection and model construction

[0102] 1.1 Data collection: Based on the specified typical office building space, parametric modeling is carried out using Rhinoceros + Grasshopper + Ladybug + Honeybee, and natural lighting environment simulation is performed. A parametric correlation is constructed between fixed variables (building geometric parameters, building material parameters), regulatory behavior variables (rotation angle of sunshade louvers), time parameters (year, month, day), climate parameters (sky model, meteorological EPW data), and the ratio of the matrix or area of indoor effective natural daylight illumination sUDI. These data are recorded during the simulation to construct the corresponding sample dataset.

[0103] 1.2 Data preprocessing: It includes data encoding and data standardization.

[0104] 1.2.1 Data encoding: If the dataset contains categorical variables (such as lighting type, space type, etc.), they need to be converted into numerical types for machine learning models to process. Specifically, season and date data need to be label-encoded (LabelEncoding). Specifically, by calling the label encoding (LabelEncoder) function in the machine learning toolkit scikit-learn, string character information is converted into numerical values.

[0105] 1.2.2 Data standardization: For input feature variables with different variation ranges, normalization and standardization of data dimensions and distributions are required to make different features have the same scale and avoid affecting the accuracy of result prediction due to data amplitude differences. Z-score standardization is performed on numerical features. Common standardization methods include Min-Max Scaling and Standard Scaling.

[0106] 1.3 Model training and validation: Using the scikit-learn library in Python, the preprocessed dataset is divided into a training set and a test set in a ratio of 8:2. Initialize the random forest regression model, set appropriate parameters (such as n_estimators = 100, max_depth = 10, etc.), and use the training set to train the model. Use the test set to validate the trained model, and calculate evaluation indicators such as mean squared error (MSE), root mean squared error (RMSE), mean absolute error (MAE), and coefficient of determination (R 2 ) to evaluate the accuracy of the model.

[0107] Step 2: Problem modeling

[0108] 2.1 State Space Definition: A state describes the current situation of the system. In the problem of regulating the visual comfort performance of intelligent shading blinds, the state includes season, date, time of day, climate parameters (direct normal irradiance outdoors, diffuse irradiance outdoors, dry bulb temperature), and the angle of the shading blind blades. These pieces of information are combined into a multi-dimensional vector as the state s. For example, s = [season, date, time of day, climate parameters, angle of the shading blind blades]. The state space S is the set of all possible state vectors.

[0109] 2.2 Action Space Definition: The adjustment angle range of the shading blinds is divided into several discrete values (such as 0°, 10°, 20°, …… 90°), and these discrete values constitute the action space A.

[0110] Step 3: Obtain the state transition function and the reward function

[0111] 3.1 State Transition Function: The trained prediction model of the visual comfort index of the light environment is encapsulated as a function. Given the current state s and action a as inputs, it outputs the predicted value of the next state s'. The state transition probability P(s'|s,a) is determined according to the prediction result.

[0112] 3.2 Reward Function: Implement the reward function R(s,a) in the code. Calculate the value function V(s') of the next state and the value function V(s) of the current state according to the state s and action a, and the difference between the two is the reward value.

[0113] Step 4: Construct an intelligent shading control agent based on dynamic programming

[0114] 4.1 Initialization: Initialize the value function V(s) to zero using a Python array, and the policy π(s) as a randomly selected action.

[0115] 4.2 Iteration of the dynamic programming algorithm

[0116] 4.2.1 Policy Evaluation: Write a function to implement the iterative update of the Bellman equation until the value function converges.

[0117] 4.2.2 Policy Improvement: Write a function to calculate the action value function Q(s,a) according to the current value function, and select the action that maximizes Q(s,a) as the new policy.

[0118] 4.2.3 Repeat the policy evaluation and policy improvement steps until the policy no longer changes.

[0119] Step 5: Implementation of intelligent shading control

[0120] 5.1 Real-time status monitoring: Before use, install devices such as light sensors and angle sensors to collect the parameters of the light environment and the current angle of the sunshade louvers in real time, and transmit the data to the control system for obtaining the status that the regulation agent needs to observe.

[0121] 5.2 Action decision-making: Input the status information collected in real time into the sunshade regulation agent based on dynamic programming, and select the adjustment angle of the sunshade louvers according to the converged optimal strategy.

[0122] 5.3 Action execution: The control system sends the selected action instruction to the driving device of the sunshade louvers to perform the angle adjustment operation.

[0123] 5.4 Loop iteration: Set a timer to repeat the above steps of real-time status monitoring, action decision-making, and action execution every certain period (such as 5 minutes) to achieve real-time dynamic regulation of the sunshade louvers.

[0124] The present invention also provides an electronic device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the steps of the intelligent sunshade louver visual comfort performance regulation method based on the dynamic programming algorithm are implemented.

[0125] The present invention also provides a computer-readable storage medium for storing computer instructions, and when the computer instructions are executed by a processor, the steps of the intelligent sunshade louver visual comfort performance regulation method based on the dynamic programming algorithm are implemented.

[0126] The memory in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read only memory (ROM), a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchlink DRAM (SLDRAM), and direct rambus RAM (DRRAM). It should be noted that the memory of the method described in the present invention is intended to include but not limited to these and any other suitable types of memory.

[0127] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center in a wired manner (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wirelessly (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer, or a data storage device such as a server or data center that includes one or more integrated available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a high-density digital video disc (DVD)), or a semiconductor medium (such as a solid state disc (SSD)), etc.

[0128] In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor or the instructions in the form of software. The steps of the method disclosed in the embodiments of the present application can be directly embodied as being executed by the hardware processor, or executed by the combination of the hardware and software modules in the processor. The software module can be located in a mature storage medium in the art such as random access memory, flash memory, read-only memory, programmable read-only memory, or electrically erasable programmable memory, register, etc. This storage medium is located in the memory, and the processor reads the information in the memory and combines its hardware to complete the steps of the above method. To avoid repetition, it will not be described in detail here.

[0129] It should be noted that the processor in the embodiments of the present application may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method embodiments may be completed by the integrated logic circuit in the hardware of the processor or instructions in software form. The above-mentioned processor may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present application may be directly embodied as being executed and completed by the hardware decoding processor, or executed and completed by a combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory, and the processor reads the information in the memory and combines its hardware to complete the steps of the above method.

[0130] The above has introduced in detail a method for regulating the visual comfort performance of intelligent shading louvers based on the dynamic programming algorithm. In this article, specific examples are used to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for controlling the visual comfort performance of intelligent sunshade blinds based on a dynamic programming algorithm, characterized in that: The method comprises the following steps: Step 1: In the selected typical office building space, using the data simulation method, collect the potential influencing factors of the light environment visual comfort performance and the visual comfort evaluation index values, and construct the office space light environment visual comfort data set; Step 2: Based on the nonlinear correlation between the potential influencing factors of the light environment visual comfort performance and the visual comfort evaluation index value, the random forest model algorithm in machine learning is trained with the data set in step 1 to complete the construction of the light environment visual comfort index prediction model; the model construction process includes data set preprocessing, test set and training set segmentation, and model training verification; Step 3, taking the visual comfort evaluation index as the optimization target, using the light environment visual comfort index prediction model of step 2, combined with the dynamic programming algorithm, develops an intelligent sunshade blinds visual comfort performance control based on the dynamic programming algorithm; specifically, taking the light environment visual comfort index prediction model of step 2 as the state model, obtaining the state transfer function and reward function, and constructing a sunshade control intelligent agent based on dynamic programming.

2. The method according to claim 1, characterized in that In the process of data collection and model building, Firstly, data on the potential influencing factors of the visual comfort performance of the light environment and the corresponding visual comfort evaluation index values ​​were collected to construct a data set; Secondly, preprocess the data set, including data cleaning, coding and standardization; Finally, the preprocessed data set is divided into training set and test set according to a certain ratio. The training set is trained using the random forest model algorithm to build a prediction model for the visual comfort index of the light environment, and the accuracy of the model is verified using the test set.

3. The method according to claim 1, characterized in that The problem modeling of the control method is specifically as follows: Define the state space: The state includes various parameters of the light environment, the current angle of the sunshade blinds and time information. The state is represented by s, and the state space S is the set of all possible states. Define action space: Action refers to adjusting the angle of the sunshade blinds. Action is represented by a. The action space A is the set of all possible actions. Determine the optimization target: The optimization target is set to maximize the visual comfort evaluation index, and the output of the light environment visual comfort index prediction model is used as the value function V(s).

4. The method according to claim 3, characterized in that State transfer function: The light environment visual comfort index prediction model is regarded as a state model. Given the current state s and action a, the predicted value of the next state s' is obtained through the model; if the state transfer is deterministic, P(s'|s,a) = 1 if and only if s' is the next state predicted by the model, otherwise P(s'|s,a) = 0.

5. The method according to claim 4, characterized in that Reward function: The reward function R(s,a) is defined as the change in the visual comfort evaluation index of state s after taking action a, that is, R(s,a) = V(s') - V(s), where s' is the next state transferred to after taking action a.

6. The method according to claim 5, characterized in that Initialization is performed in the process of constructing the sunshade control agent based on dynamic programming. The value function is initialized to zero and is applicable to all states. The value function represents the expected value of the long-term cumulative reward that can be obtained by following the optimal strategy starting from the initial state. The initialization strategy is a random strategy, which randomly selects an action for each state. Specifically, the value function V(s) is initialized to zero and is applicable to all states s∈S. Initialize the initialization strategy π(s) as a random strategy, that is, for each state s, randomly select an action a∈A.

7. The method according to claim 6, characterized in that In the process of constructing the sunshade control intelligent agent based on dynamic programming, the dynamic programming algorithm is iterated: the value function is calculated according to the current strategy, and the value function is updated for each state using the Bellman equation; when improving the strategy, the strategy is improved according to the current value function; For each state, select the action that maximizes the action-value function as the new policy; Repeat the policy evaluation and policy improvement steps until the policy converges, i.e. the policy no longer changes.

8. The method according to claim 1, characterized in that The specific process of realizing intelligent sunshade control is as follows: Real-time status monitoring: Use sensors to monitor the parameters of the light environment and the current status of the sunshade blinds in real time, and input this information as the current state s into the sunshade control agent based on dynamic programming; Action decision: Based on the current state s, the agent selects an action a through the converged optimal strategy π(s), that is, determines the optimal adjustment angle of the sunshade blinds; Action execution: Send the selected action a to the sunshade louver control system to perform the angle adjustment operation of the sunshade louver; Iteration: Repeat the real-time status monitoring, action decision-making and action execution steps to achieve real-time dynamic control of the sunshade blinds.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium for storing computer instructions, characterized in that: When the computer instructions are executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images