Indoor space environment control method based on optimization algorithm and related equipment

Abstract

The invention provides an indoor space environment control method based on an optimization algorithm and related equipment. The method comprises the following steps: constructing an indoor comprehensive environment model in combination with historical equipment control parameters and historical space environment data; constructing a comprehensive objective function; creating a state space of the ant colony optimization algorithm, and generating constraint conditions of the ant colony optimization algorithm; taking the comprehensive target function as an algorithm optimization target, and generating an initial equipment control strategy of the environment control equipment by adopting an ant colony optimization algorithm based on the indoor comprehensive environment model; taking the optimal parameter interval as a strategy initial solution of a model prediction control strategy, and generating a strategy prediction result through the model prediction control strategy based on the indoor comprehensive environment model; and updating algorithm parameters of the ant colony optimization algorithm according to the strategy prediction result, and generating an optimal equipment control strategy of the environment control equipment. The method has the effect of meeting the environmental experience comfort level of the user on the basis of the lowest energy consumption.

Description

Technical Field

[0001] The present invention belongs to the technical field of environmental control optimization, and in particular relates to an indoor space environment control method based on an optimization algorithm and related equipment. Background Art

[0002] In the face of global climate change and energy crisis, improving building energy efficiency has become an urgent task. In this context, application scenarios such as intelligent building management systems, home automation control, and office environment optimization are becoming increasingly popular. These applications aim to achieve intelligent adjustment of indoor environmental parameters through advanced sensing technology, network communication, and control algorithms to create a comfortable, healthy, and efficient indoor environment.

[0003] However, the indoor environment is a highly complex and dynamically changing system involving multiple factors that influence each other. Although traditional environmental control equipment can be adjusted intelligently based on environmental data, it often uses simple threshold control or fixed control logic, lacking the ability to accurately grasp and predict the dynamic changes of the indoor environment. In order to meet users' requirements for comfort, the system tends to over-adjust, resulting in a large amount of energy consumption; and when switching to energy-saving mode, it is often difficult to achieve the ideal comfort level, which affects the user experience. This contradiction between comfort and energy efficiency has become a major problem that plagues the industry. Summary of the invention

[0004] The present invention provides an indoor space environment control method based on an optimization algorithm and related equipment to solve the problem that it is difficult for environmental control equipment to balance the contradiction between comfort and energy efficiency.

[0005] In a first aspect, the present invention provides an indoor space environment control method based on an optimization algorithm, the method comprising the following steps:

[0006] Acquire historical device control parameters of the environment control device and historical space environment data corresponding to the historical device control parameters, wherein the historical space environment data includes historical indoor environment data inside the indoor closed space where the environment control device is deployed and historical outdoor environment data outside the indoor closed space;

[0007] Building an indoor comprehensive environment model of the indoor enclosed space by combining the historical equipment control parameters and the historical space environment data;

[0008] Constructing a comprehensive objective function according to the indoor environmental comfort of the indoor enclosed space and the equipment energy consumption index of the environmental control equipment;

[0009] Creating a state space of an ant colony optimization algorithm based on the device control parameter list of the environmental control device, and generating constraint conditions of the ant colony optimization algorithm according to parameter limits of each of the device control parameters in the device control parameter list;

[0010] Taking the comprehensive objective function as the algorithm optimization target, generating an initial device control strategy for the environmental control device based on the indoor comprehensive environment model and using the ant colony optimization algorithm, wherein the initial device control strategy includes an optimal parameter range for each of the device control parameters;

[0011] Using the optimal parameter interval as the initial solution of the model predictive control strategy, using the comprehensive objective function as the strategy optimization target of the model predictive control strategy, and generating a strategy prediction result based on the indoor comprehensive environment model and through the model predictive control strategy;

[0012] The algorithm parameters of the ant colony optimization algorithm are updated according to the strategy prediction result, and the updated ant colony optimization algorithm is used to generate the optimal device control strategy of the environmental control device.

[0013] Optionally, the step of constructing the indoor comprehensive environment model of the indoor closed space by combining the historical device control parameters and the historical space environment data comprises the following steps:

[0014] Constructing a data association driving model between the historical equipment control parameters and the historical spatial environment data based on a convolutional neural network and a long short-term memory network;

[0015] Building an indoor space physical model of the indoor enclosed space in combination with the historical indoor environment data and the historical outdoor environment data, wherein the indoor space physical model includes a thermodynamic model, a humidity model, an air quality model, and a lighting model;

[0016] The data association driven model and the indoor space physical model are integrated into an indoor comprehensive environment model of the indoor enclosed space by using a gray box model method.

[0017] Optionally, the formula of the thermodynamic model is as follows:

[0018]

[0019] Where: T in represents the indoor ambient temperature of the indoor enclosed space, C represents the heat capacity of the indoor air, Q H represents the heat provided by the environmental control equipment, Q s Indicates outdoor solar radiation heat, Q in represents the heat generated by the indoor heat source, UA represents the total heat transfer coefficient of the building where the indoor enclosed space is located, Tout Indicates the outdoor ambient temperature;

[0020] The formula of the humidity model is as follows:

[0021]

[0022] Where: W in represents the indoor absolute humidity of the indoor enclosed space, V represents the indoor volume of the indoor enclosed space, m s represents the air supply coverage of the environmental control equipment, W s Indicates the absolute humidity of the supply air, m r Indicates the return air volume of the environmental control equipment, m in Indicates the amount of infiltrated air, W out represents the outdoor absolute humidity, and G represents the indoor humidity generation rate;

[0023] The formula of the air quality model is as follows:

[0024]

[0025] Where: C in represents the CO2 concentration in the indoor enclosed space, G CO2 represents the CO2 generation rate of the indoor enclosed space, Q represents the ventilation volume of the indoor enclosed space, C out Indicates outdoor CO2 concentration;

[0026] The formula of the lighting model is as follows:

[0027] E=k1L+k2I s +δ

[0028] Where: E represents the indoor illumination of the indoor enclosed space, L represents the brightness of the artificial lighting controlled by the environmental control device, I s represents outdoor solar irradiance, k1 and k2 represent equipment lighting influence coefficient and natural lighting influence coefficient respectively, and δ represents the adjustment coefficient.

[0029] Optionally, the step of fusing the data association driven model and the indoor space physical model into an indoor comprehensive environment model of the indoor enclosed space by using a gray box model method comprises the following steps:

[0030] For any of the indoor space physical models, a new model error term is introduced into the indoor space physical model;

[0031] Combining the data association driven model with the indoor space physical model to obtain an initial indoor comprehensive environment model of the indoor enclosed space;

[0032] After preprocessing the historical equipment control parameters and the historical space environment data, the historical equipment control parameters and the historical space environment data are integrated into a model training set;

[0033] Inputting the model training set into the initial indoor comprehensive environment model, outputting the environment association change result based on the historical equipment control parameters and through the data association driving model, combining the environment association change result and the historical space environment data and outputting the indoor space environment prediction data through the indoor space physical model;

[0034] The mean square error between the indoor space environment prediction data and the historical indoor environment data is used as a model loss function;

[0035] Simultaneously adjusting the model error term of the indoor space physical model and the model parameters of the data association driven model according to the model loss function;

[0036] The initial indoor comprehensive environment model is continuously trained using the model training set until the model loss function reaches a minimum value, thereby obtaining a trained indoor comprehensive environment model.

[0037] Optionally, after preprocessing the historical device control parameters and the historical space environment data, integrating the historical device control parameters and the historical space environment data into a model training set comprises the following steps:

[0038] Unifying the data sampling rates of the historical equipment control parameters and the historical spatial environment data;

[0039] Align the data of the historical equipment control parameters in each time node with the data of the historical spatial environment data in the previous time node;

[0040] The historical equipment control parameters and the historical spatial environment data after data alignment are integrated into a model training set.

[0041] Optionally, the step of using the comprehensive objective function as an algorithm optimization target, generating an initial device control strategy for the environment control device based on the indoor comprehensive environment model and using the ant colony optimization algorithm comprises the following steps:

[0042] Initializing algorithm parameters of the ant colony optimization algorithm;

[0043] Randomly generate ant positions of a plurality of ants in the state space based on the constraint condition, where different ant positions represent different device control parameters;

[0044] For each of the ant positions, the corresponding comprehensive objective function value is calculated using the indoor comprehensive environment model;

[0045] Updating the pheromone concentration in the state space, and updating the ant position in combination with the comprehensive objective function value and the updated pheromone concentration;

[0046] Repeat the above steps of calculating the comprehensive objective function value and updating the ant positions until the maximum number of iterations of the ant colony optimization algorithm is reached, and output the optimal ant path under the constraint conditions;

[0047] The optimal parameter interval of each device control parameter of the environmental control device is determined according to the optimal ant path, and the optimal parameter interval of each device control parameter of the environmental control device is integrated into an initial device control strategy.

[0048] Optionally, the step of using the optimal parameter interval as the initial solution of the model predictive control strategy, using the comprehensive objective function as the strategy optimization target of the model predictive control strategy, and generating a strategy prediction result based on the indoor comprehensive environment model and through the model predictive control strategy comprises the following steps:

[0049] Using the optimal parameter interval as the strategy initial solution of the model predictive control strategy;

[0050] Determining the prediction time domain and control time domain of the model prediction control strategy based on the initial solution of the strategy and through the indoor comprehensive environment model;

[0051] On the basis of the prediction time domain, the comprehensive objective function is used as the strategy optimization target of the model predictive control strategy, and the control sequence in the control time domain is rollingly optimized using the sequential quadratic programming method until the comprehensive objective function takes the minimum value to obtain the optimal control sequence, and the optimal control sequence is used as the strategy prediction result generated by the model predictive control strategy.

[0052] Optionally, updating the algorithm parameters of the ant colony optimization algorithm according to the strategy prediction result, and using the updated ant colony optimization algorithm to generate the optimal device control strategy for the environmental control device includes the following steps:

[0053] Updating the pheromone distribution of the ant colony optimization algorithm according to the prediction result of the strategy;

[0054] Generating target constraints of the ant colony optimization algorithm according to the optimal parameter interval;

[0055] The comprehensive objective function is used as the algorithm optimization target, the state space and the target constraint conditions are combined and the ant colony optimization algorithm is used to generate the optimal device control strategy of the environmental control device.

[0056] In a second aspect, the present invention also provides an indoor space environment control system based on an optimization algorithm, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the indoor space environment control method based on the optimization algorithm as described in the first aspect is implemented.

[0057] In a third aspect, the present invention further provides an environment control device, including the indoor space environment control system based on the optimization algorithm described in the second aspect.

[0058] The beneficial effects of the present invention are:

[0059] First, the present invention can more accurately capture the dynamic change characteristics of the indoor environment. By constructing a comprehensive environmental model, it fully considers the complex interaction of indoor and outdoor environmental factors, and improves the accuracy and adaptability of environmental control. Secondly, the present invention effectively solves the contradiction between comfort and energy saving in traditional methods by incorporating comfort and energy efficiency into a unified optimization framework, and achieves balanced optimization of the two. The present invention uses an ant colony optimization algorithm to generate an initial control strategy, and combines it with a model predictive control strategy for further optimization, so that the optimization algorithm parameters can be continuously updated according to the strategy prediction results, so that the control strategy can be continuously improved over time. This multi-level optimization method greatly improves the efficiency and effectiveness of the control strategy. Moreover, the control strategy of the present invention takes into account the limit constraints of the equipment control parameters, ensuring that the generated control strategy is within the actual operable range, improving the practicality and reliability of the solution, thereby minimizing energy consumption while ensuring user comfort, and providing an efficient, accurate and learning-capable environmental control solution for the fields of intelligent buildings and energy conservation and environmental protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is a flow chart of an indoor space environment control method based on an optimization algorithm in one embodiment of the present application.

[0061] Figure 2 This is a schematic diagram of the system structure of an environmental control device in one embodiment of the present application. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the present application belong to the scope of protection of this application.

[0063] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0064] Figure 1 FIG. 1 is a flow chart of an indoor space environment control method based on an optimization algorithm in one embodiment. It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps in the above method may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps. Figure 1 As shown, the indoor space environment control method based on the optimization algorithm disclosed in the present invention specifically includes the following steps:

[0065] S101. Obtain historical device control parameters of an environment control device, and historical space environment data corresponding to the historical device control parameters.

[0066] Among them, the historical space environment data includes the historical indoor environment data inside the indoor enclosed space where the environment control equipment is deployed and the historical outdoor environment data outside the indoor enclosed space. Specifically, the historical equipment control parameters of the environment control equipment are first collected. These parameters include temperature setting value, humidity setting value, wind speed setting value, and brightness setting value of the lighting system. At the same time, it is also necessary to collect the historical space environment data corresponding to these control parameters. These environmental data are divided into two parts: indoor environment data and outdoor environment data. Indoor environment data includes parameters such as indoor temperature, humidity, CO2 concentration, and illumination. These data are usually collected by various sensors installed indoors. For example, temperature and humidity sensors can monitor indoor temperature and humidity changes in real time; CO2 sensors can detect indoor air quality; and illumination sensors are used to measure indoor light intensity. These sensors usually record data at a certain time interval (such as every 5 minutes or every 10 minutes) to form a continuous time series.

[0067] Outdoor environmental data includes parameters such as outdoor temperature, humidity, wind speed, wind direction, and sunshine intensity. These data can be obtained through weather stations installed outside buildings or from local meteorological departments. The frequency of outdoor environmental data collection is usually consistent with that of indoor environmental data to facilitate subsequent data analysis and modeling. In actual operation, a database can be established to store these historical data. The structure of the database includes the following main tables:

[0068] Equipment control parameter table: records the set values ​​of each environmental control device at each time point.

[0069] Indoor environment data table: records the indoor environment parameters at each time point.

[0070] Outdoor environment data table: records the outdoor environment parameters at each time point.

[0071] These tables are linked by timestamps to ensure data consistency and traceability. The time span of data collection usually needs to cover at least one complete seasonal cycle, preferably a full year, in order to capture the laws of environmental changes under different seasons and weather conditions. The frequency of data collection needs to be high enough to capture the dynamic characteristics of environmental changes, but also take into account the limitations of storage and processing capabilities.

[0072] Data preprocessing is an important part of this step. First, the raw data needs to be cleaned to remove obvious outliers and missing values. Outliers may be caused by sensor failure or temporary interference, which can be identified and processed by setting reasonable thresholds or using statistical methods (such as the 3σ rule). For missing values, interpolation methods can be used to fill them, such as linear interpolation or more complex time series interpolation methods. In addition, the data needs to be standardized to unify data of different scales to the same dimension. Commonly used standardization methods include minimum-maximum standardization and Z-score standardization. Through these processes, a clean, consistent historical data set that can be used for subsequent analysis can be obtained. This data set will provide an important foundation for subsequent modeling and optimization, enabling environmental control equipment to make decisions and control more intelligently and efficiently based on the patterns and laws of historical data.

[0073] S102. Construct an indoor comprehensive environment model of the indoor enclosed space by combining historical equipment control parameters and historical space environment data.

[0074] Among them, in this step, the main task is to build an indoor comprehensive environment model of the indoor enclosed space. The core goal of this model is to establish a data association relationship between the control parameters of the environmental control equipment and the indoor and outdoor environmental data. This association relationship will help understand and predict the impact of changes in control parameters on the indoor environment, and provide a theoretical basis for subsequent optimization control. In order to build an indoor comprehensive environment model, it is necessary to use the historical data collected in step S101. A variety of modeling methods can be used. For example, the physical model method can be based on physical equations and use optimization algorithms such as least squares to estimate model parameters. The data-driven method uses machine learning algorithms, such as multivariate linear regression, support vector machine regression, random forest or neural network, to directly learn the relationship between control parameters and environmental parameters from data.

[0075] The constructed indoor comprehensive environmental model will provide important support for subsequent optimization control. It can help predict indoor environmental changes under different control strategies, evaluate the effects of control strategies, and provide a simulation platform for finding the optimal control strategy. At the same time, this model can also be used for real-time monitoring and anomaly detection. By comparing the model prediction value and the actual measurement value, the anomaly or performance degradation of the environmental control system can be discovered in time.

[0076] S103. Construct a comprehensive objective function based on the indoor environmental comfort of the indoor enclosed space and the equipment energy consumption index of the environmental control equipment.

[0077] In this step, a comprehensive objective function needs to be constructed, which will take into account both the indoor environmental comfort and the energy consumption index of the environmental control equipment. This comprehensive objective function will serve as the common optimization target of the subsequent ant colony optimization algorithm and model predictive control strategy to minimize energy consumption while ensuring the indoor environmental comfort. The process of constructing a comprehensive objective function usually includes the following key aspects:

[0078] 1. Indoor environment comfort evaluation:

[0079] Indoor environmental comfort usually includes multiple aspects such as thermal comfort, humidity comfort, air quality comfort and visual comfort. A comfort index can be defined for each aspect and then combined into a comprehensive comfort index.

[0080] a) Thermal comfort: PMV (predicted mean vote) index or adaptive thermal comfort model can be used. The calculation formula of PMV model is as follows:

[0081] PMV=(0.303e -0.036M +0.028)L

[0082] Where M is the human metabolic rate and L is the heat load. PMV values ​​between -0.5 and +0.5 are considered comfortable.

[0083] b) Humidity comfort: It can be evaluated directly using relative humidity or combined with temperature using dew point temperature. The ideal relative humidity range is usually between 40% and 60%.

[0084] c) Air quality comfort: C02 concentration can be used as an indicator (other gas concentrations that affect indoor air quality can also be used as indicators). A C02 concentration below 1000ppm is generally considered to be good indoor air quality.

[0085] d) Visual comfort: It can be evaluated using the indoor illumination level. Different scenes have different recommended illumination ranges. For example, the recommended illumination range for offices is usually between 300 and 500 lux.

[0086] The comprehensive comfort index can be expressed as:

[0087] P C =w1f(PMV)+w2g(RH)+w3h(CO2)+w4i(Lux)

[0088] Among them, f(), g(), h(), i() are the normalized functions of each comfort index, and w1, w2, w3, w4 are the weights of each index.

[0089] 2.Energy consumption index evaluation:

[0090] The energy consumption index usually includes the energy consumption of air conditioning system, lighting system and ventilation system in environmental control equipment. The total energy consumption can be calculated using the following formula:

[0091] P E =E H +E L +E V

[0092] Among them, E H is the energy consumption of the air conditioning system, E L is the energy consumption of the lighting system, E V is the energy consumption of the ventilation system.

[0093] The energy consumption of each system can be estimated by the power and operating time of the equipment, for example:

[0094]

[0095] Among them, P i is the rated power of the ith air conditioner, t i is the running time, COP i It is the energy efficiency ratio.

[0096] 3. Construction of comprehensive objective function:

[0097] The comprehensive objective function needs to consider both comfort and energy consumption, which can be in the form of a weighted sum:

[0098] J=α·P C -β·P E

[0099] Among them, α and β are weight coefficients used to balance the importance of comfort and energy consumption. The negative sign is used here because we want to maximize comfort while minimizing energy consumption.

[0100] 4. Definition of constraints:

[0101] When constructing the objective function, some constraints need to be defined to ensure that the optimization results are within the practical range. These constraints include:

[0102] a) Range limitations of equipment operating parameters, such as the range of temperature setting values, the range of wind speed, etc.

[0103] b) Minimum requirements for comfort indicators, such as the absolute value of PMV not exceeding 0.5.

[0104] c) An upper limit on energy consumption, determined by the building’s energy quota or the capacity of the electrical system.

[0105] These constraints can be expressed as inequality constraints.

[0106] 5. Normalization of objective function:

[0107] Due to the different dimensions and magnitudes of the comfort index and the energy consumption index in the normalization of the objective function, direct addition will cause one aspect to dominate the optimization process. Therefore, these indicators need to be normalized to keep their numerical ranges consistent. Commonly used normalization methods include minimum-maximum normalization and Z-score normalization. Through normalization, it can be ensured that comfort and energy consumption are balanced in the optimization process. In the process of constructing a comprehensive objective function, optimization objectives of different time scales also need to be considered. Short-term goals focus more on immediate comfort, while long-term goals focus more on energy efficiency and equipment life. The short-term and long-term goals can be balanced by introducing a time weighting factor, so that the control strategy can strike a balance between immediate comfort and long-term benefits.

[0108] In addition, the construction of the objective function should also take into account the dynamic characteristics of the environmental control system. For example, there will be large fluctuations in energy consumption during the start and stop of the air conditioning system, and frequent switching will affect the life of the equipment. Therefore, a penalty term can be added to the objective function to reduce unnecessary frequent adjustments. In practical applications, it is necessary to continuously adjust the weights and constraints of various indicators based on system operation data and user feedback to achieve the best control effect. The objective function form that best suits the specific environment and user needs is determined through repeated trials and optimization.

[0109] S104. Create a state space of the ant colony optimization algorithm based on the device control parameter list of the environmental control device, and generate constraint conditions of the ant colony optimization algorithm according to parameter limits of each device control parameter in the device control parameter list.

[0110] Among them, this step provides the search space and search boundary for the subsequent ant colony optimization algorithm, and is an important basis for the optimization process. First, it is necessary to determine the list of control parameters of the environmental control equipment. This list usually includes the temperature setting value, humidity setting value, wind speed setting value of the air-conditioning system, the brightness setting value of the lighting system, the air change rate setting value of the ventilation system, etc. Each parameter represents a control dimension, and together they constitute a multidimensional control space. The process of creating a state space is actually the process of defining this multidimensional control space. Each control parameter corresponds to a dimension in the space, and the range of the parameter value defines the range of the dimension. For example, if the temperature setting value ranges from 18°C ​​to 30°C, then in the temperature dimension, the state space extends from 18 to 30.

[0111] For discrete parameters, such as wind speed with three gears of low, medium and high, they can be mapped to the numerical space, for example, 1, 2, and 3 are used to represent these three gears. The purpose of this is to unify all parameters into the numerical space to facilitate subsequent optimization calculations. The definition of the state space is usually in the form of a vector. Assuming there are n control parameters, a point in the state space can be represented as an n-dimensional vector, with each component corresponding to the value of a control parameter. This vector represents a control strategy for the environmental control system.

[0112] Next, constraints need to be generated based on the parameter limits of the device control parameters. These constraints define the legal search area in the state space. Constraints usually include the following categories:

[0113] 1. Parameter range constraints: Each control parameter has a reasonable range of values. Exceeding this range will cause the device to not work properly or create an uncomfortable environment. For example, the indoor temperature setting value is limited to between 18℃ and 30℃.

[0114] 2. Correlation constraints between parameters: Some parameters have mutual influence or correlation. For example, the set values ​​of temperature and humidity need to be considered at the same time to avoid too humid or too dry conditions.

[0115] 3. Equipment capacity constraints: The actual capacity of the environmental control equipment will also constrain the parameter settings. For example, the cooling and heating capacity of the air-conditioning system is limited, which limits the indoor temperature range that can be achieved under extreme external environments.

[0116] 4. Energy consumption constraints: An upper limit on total energy consumption needs to be set to meet energy-saving requirements or capacity limitations of the power system.

[0117] 5. Comfort constraints: Minimum comfort requirements need to be set to ensure that the optimization results do not sacrifice user experience.

[0118] These constraints play an important role in the ant colony optimization algorithm. They limit the range of movement of ants in the state space and ensure that every solution considered is feasible. In the algorithm implementation, these constraints are usually converted into judgment conditions or penalty functions to guide the path selection of ants and the update of pheromones. Through this step, a structured search space and clear constraints are defined for the ant colony optimization algorithm. This provides clear boundaries and guidance for the subsequent optimization process, helping the algorithm to more efficiently find the optimal control strategy that meets all requirements.

[0119] S105. Taking the comprehensive objective function as the algorithm optimization target, an initial device control strategy for the environmental control device is generated based on the indoor comprehensive environment model and using the ant colony optimization algorithm.

[0120] Among them, the ant colony optimization algorithm is a swarm intelligence optimization algorithm inspired by the foraging behavior of ants. In this algorithm, when ants move in the state space, they choose the path based on the pheromone concentration and heuristic information. The pheromone concentration reflects the historical search experience, while the heuristic information represents the estimated distance from the current state to the target. The specific implementation process of the algorithm is as follows:

[0121] A certain number of ants are randomly generated in the state space, each ant represents a set of initial device control parameters. At the same time, the pheromone distribution is initialized, and the initial pheromone concentration on all paths is usually set to the same small positive number.

[0122] Each ant chooses the next direction to move based on the pheromone concentration and heuristic information at the current location. In the environmental control problem, this is equivalent to adjusting the values ​​of various control parameters. The probability of selection is proportional to the weighted product of the pheromone concentration and the heuristic information. The pheromone concentration reflects the historical search experience, and the heuristic information can be designed based on the estimated contribution of the current parameter setting to the objective function.

[0123] The solution (i.e., a set of control parameters) constructed by each ant is evaluated. The evaluation criteria are the comprehensive objective function defined in step S103, and the constraints set in step S104 are also considered. If a solution violates the constraints, its score can be reduced by using a penalty function.

[0124] The pheromone distribution in the state space is updated according to the evaluation results of each ant. Generally, the paths with higher scores will get more pheromone increase. At the same time, the pheromone on all paths will evaporate to a certain extent to avoid the algorithm converging to the local optimal solution too early.

[0125] Repeatedly update the ant movement position, evaluate the solution constructed by the ants, and update the pheromone distribution until the preset number of iterations is reached or a certain convergence condition is met.

[0126] After the algorithm is finished, the optimal solution is selected as the initial equipment control strategy. This strategy contains the optimal interval of each control parameter.

[0127] In practical applications, multiple independent optimization runs are required to reduce the impact of random factors and improve the reliability of the results. In addition, improved strategies such as elite strategies and local search strategies can be introduced to improve the performance of the algorithm. Through the ant colony optimization algorithm, a near-optimal control strategy can be efficiently searched in a complex multi-dimensional state space. This initial control strategy provides a good starting point for subsequent model predictive control, which helps to improve the efficiency and performance of the entire control system.

[0128] It should be noted that the ant colony optimization algorithm generates an initial control strategy, which is not enough to cope with the complex dynamic changes of the environment. This is why it is necessary to introduce model predictive control in the subsequent steps to achieve real-time response and adjustment to environmental changes.

[0129] S106. Use the optimal parameter interval as the initial solution of the model predictive control strategy, use the comprehensive objective function as the strategy optimization target of the model predictive control strategy, and generate a strategy prediction result based on the indoor comprehensive environment model and through the model predictive control strategy.

[0130] In this step, the optimal parameter interval obtained by the ant colony optimization algorithm in step S105 is used as the initial solution of the model predictive control (MPC) strategy, and the comprehensive objective function defined in step S103 is used as the optimization target of MPC. Based on the indoor comprehensive environment model constructed in step S2, the strategy prediction result is generated by the MPC strategy. This process combines the global search capability of ant colony optimization with the dynamic prediction and real-time optimization capabilities of MPC to achieve more accurate and adaptable environmental control.

[0131] Model predictive control is an advanced control strategy that uses system models to predict system behavior in the future and optimizes the control sequence based on this. The core idea of ​​MPC is to perform an optimization calculation in each control cycle, but only execute the first step of the optimization result, and then restart the optimization process in the next control cycle. This rolling optimization method enables MPC to adapt to the dynamic changes and uncertainties of the system.

[0132] In practical applications, the indoor comprehensive environmental model needs to be updated regularly to adapt to the slow changes in system characteristics (such as changes in building thermal characteristics caused by seasonal changes). This can be achieved through online parameter estimation or periodic model recalibration. Given the existence of model errors and environmental disturbances, the MPC strategy needs to have a certain degree of robustness. This can be achieved by considering uncertainty in the optimization process, such as using robust MPC or stochastic MPC methods. These methods can consider the uncertainty range of model parameters or the statistical characteristics of environmental disturbances during the optimization process, thereby generating more robust control strategies. MPC needs to complete the optimization calculation within each control cycle, so computational efficiency is an important consideration. Fast MPC algorithms, such as explicit MPC or approximate dynamic programming, can be used to increase the computational speed. In addition, the length of the control cycle can be adjusted according to the dynamic characteristics of the system to strike a balance between control performance and computational load. During the optimization process, various constraints need to be strictly observed, including equipment physical limitations, comfort requirements, and energy consumption limitations. Soft constraint methods can be used, that is, the degree of constraint violation is incorporated into the objective function to balance control performance and constraint satisfaction.

[0133] S107. Update the algorithm parameters of the ant colony optimization algorithm according to the strategy prediction result, and use the updated ant colony optimization algorithm to generate the optimal device control strategy for the environmental control device.

[0134] Among them, the pheromone distribution of the ant colony optimization algorithm is first updated according to the prediction results of the MPC strategy. This update process reflects the MPC's prediction information on future environmental changes, which helps to guide the ant colony algorithm to search for more promising solution spaces. Specifically, for control sequences that perform well in the MPC prediction results (i.e., sequences that can maintain high comfort and low energy consumption during the prediction period), the pheromone concentration is increased in the corresponding parameter space area. For control sequences that can better meet various constraints, a higher pheromone increase is given. For sequences that have small fluctuations during the prediction period and can maintain a stable control effect, their pheromone concentration is increased. For control strategies that perform well in different prediction scenarios, higher pheromone rewards are given.

[0135] Next, the new target constraint of the ant colony optimization algorithm is generated based on the optimal parameter interval obtained by MPC. Based on the prediction results of MPC, the maximum change rate allowed for the control parameters can be set to avoid drastic fluctuations in the control strategy. According to the comfort change trend predicted by MPC, a more stringent lower limit constraint for comfort is set. Based on the energy consumption prediction of MPC, the energy consumption constraint needs to be adjusted to ensure long-term energy efficiency. According to the prediction of MPC, some dynamic constraints need to be added, such as special requirements within a specific period of time.

[0136] Then, the ant colony optimization algorithm is run again, with the comprehensive objective function as the optimization target, and the search is performed in the updated state space while complying with the new constraints. This optimization process is similar to step S105, as follows:

[0137] The optimal control sequence obtained by MPC is used as the initial position of some ants to accelerate convergence.

[0138] The calculation method of heuristic information can be adjusted according to the prediction results of MPC to better reflect the long-term control effect.

[0139] After each iteration, MPC can be used to locally optimize the current optimal solution to further improve the quality of the solution.

[0140] Considering the multiple performance indicators predicted by MPC, multi-objective optimization methods such as Pareto optimization are needed.

[0141] During the optimization process, algorithm parameters such as pheromone renewal rate, evaporation rate, etc. can be dynamically adjusted according to the predictions of MPC.

[0142] Through this optimization process, the optimal control strategy for the environmental control equipment is finally obtained. This strategy not only takes into account the immediate control effect, but also takes into account the prediction of future environmental changes, so it can achieve better long-term control performance. During actual operation, the system will continuously monitor environmental parameters and energy consumption data, and regularly evaluate the control effect. If it is found that the actual effect deviates significantly from the expected, or the environmental conditions change significantly, the entire optimization process needs to be re-executed to maintain the optimality of the control strategy. Through this method of combining ant colony optimization algorithm, MPC and dynamic feedback, an intelligent, efficient and adaptable indoor environmental control system can be realized, which can not only ensure the comfort of users, but also achieve efficient use of energy.

[0143] In one embodiment, combining historical equipment control parameters and historical space environment data to construct an indoor comprehensive environment model of an indoor closed space includes the following steps:

[0144] A data association driven model between historical equipment control parameters and historical spatial environment data is constructed based on convolutional neural networks and long short-term memory networks;

[0145] Combining historical indoor environmental data and historical outdoor environmental data to build an indoor space physical model of the indoor enclosed space, the indoor space physical model includes a thermodynamic model, a humidity model, an air quality model and a lighting model;

[0146] The gray box model method is used to integrate the data association driven model and the indoor space physical model into an indoor comprehensive environmental model of the indoor enclosed space.

[0147] In this embodiment, a data association driven model between historical equipment control parameters and historical spatial environment data is constructed based on convolutional neural networks (CNN) and long short-term memory networks (LSTM). The core of this step is to use deep learning technology to capture the complex nonlinear relationship between equipment control parameters and indoor environmental states. First, the architecture of the CNN-LSTM hybrid model is designed, with the CNN part used to extract spatial features and the LSTM part used to capture time series characteristics. The input of CNN is a multi-channel image representing the indoor environmental state, with each channel corresponding to an environmental parameter. The typical structure of CNN includes multiple convolutional layers, activation functions (such as ReLU), and pooling layers. For example, a 3x3 convolution kernel can be used with a step size of 1 and a padding of 1 to maintain the spatial size of the feature map. The pooling layer can use a 2x2 maximum pooling with a step size of 2 to reduce the size of the feature map.

[0148] The specific CNN structure is as follows:

[0149] Convolutional layer 1: 32 3x3 convolution kernels, ReLU activation;

[0150] Max pooling layer 1: 2x2 pooling kernel, stride 2;

[0151] Convolutional layer 2: 64 3x3 convolution kernels, ReLU activation;

[0152] Max pooling layer 2: 2x2 pooling kernel, stride 2;

[0153] Convolutional layer 3: 128 3x3 convolution kernels, ReLU activation;

[0154] Global average pooling layer.

[0155] The feature map output by CNN is flattened and input into the LSTM network. The LSTM network is used to process time series data. Its core is the LSTM unit, which contains an input gate, a forget gate, and an output gate. The mathematical expression of LSTM is as follows:

[0156] Input gate: i t =σ(W i ·[h t -1,x t ]+b i )

[0157] Forget gate: f t =σ(W f ·[h t -1,x t ]+b f )

[0158] Output gate: o t =σ(W o ·[h t -1,x t ]+b o )

[0159] Candidate memory unit: C t =tanh(W C ·[h t -1,x t ]+b C )

[0160] Memory unit update: C t =f t *C t -1+i t *C t

[0161] Hidden state output: h t =o t *tanh(C t )

[0162] Among them, σ is the sigmoid function, * represents the element-wise multiplication, and W and b are learnable parameters. The output of LSTM is mapped to the final prediction result through a fully connected layer. The training of the entire model uses the backpropagation algorithm, and the loss function can choose the mean square error (MSE) or the mean absolute error (MAE). The optimizer can use the Adam algorithm, the learning rate can be set to 0.001, and the learning rate decay strategy is used. To prevent overfitting, a Dropout layer can be added before the fully connected layer, and the dropout rate is set to 0.5. In order to improve the generalization ability of the model, k-fold cross validation can be used, and k is usually 5 or 10. In addition, the Early Stopping method can be used to prevent overfitting, that is, stop training when the performance of the validation set no longer improves. The advantage of this CNN-LSTM hybrid model is that it can capture the characteristics of both spatial and temporal dimensions. CNN effectively extracts spatial distribution features, while LSTM captures the long-term dependencies of time series. This structure is particularly suitable for dealing with indoor environmental control problems because indoor environmental parameters have both spatial distribution characteristics (such as temperature gradient) and obvious time series characteristics (such as daily and weekly changes).

[0163] The data association driven model constructed by this method can learn complex nonlinear relationships and predict the indoor environmental state under different control parameters. The output of this model can be the prediction of indoor environmental parameters in the future, such as temperature, humidity, C02 concentration, etc. The performance of the model can be measured by the root mean square error (RMSE), mean absolute error (MAE) and determination coefficient (R 2 The advantage of this data-driven approach is that it does not require explicit modeling of the physical process and can automatically learn complex system dynamics.

[0164] Next, we combine historical indoor environmental data with historical outdoor environmental data to build a physical model of the indoor space of the indoor enclosed space. This physical model includes four main components: thermal dynamics model, humidity model, air quality model and lighting model. This model based on physical principles can provide a deep understanding of the mechanism of indoor environmental changes and can still provide a reliable description when data is insufficient or in the face of new situations.

[0165] First, the thermodynamic model describes the indoor temperature distribution and changes based on the energy balance principle. This model takes into account multiple heat transfer pathways, including heat conduction through walls, radiant heat through windows, cooling / heating of air conditioning systems, and indoor heat sources (such as human bodies and electrical appliances). The core equation of the thermodynamic model can be expressed as:

[0166]

[0167] Where: T inrepresents the indoor ambient temperature of the indoor enclosed space, C represents the heat capacity of the indoor air, Q H Indicates the heat provided by the environmental control equipment, Q s Indicates outdoor solar radiation heat, Q in It represents the heat generated by the indoor heat source, UA represents the total heat transfer coefficient of the building where the indoor enclosed space is located, T out represents the outdoor ambient temperature. This equation needs to be solved by numerical methods (such as the Runge-Kutta method). In order to improve the accuracy of the model, the heat capacity effect of the wall can also be considered, and the wall can be divided into multiple layers, each with its own heat balance equation.

[0168] The humidity model describes the changes in the moisture content in indoor air. This model needs to take into account the dehumidification / humidification effect of the air conditioning system, the moisture produced by human breathing and sweating, and the moisture brought by outdoor air infiltration. The basic equation of the humidity model can be expressed as:

[0169]

[0170] Where: W in represents the indoor absolute humidity of the indoor enclosed space, V represents the indoor volume of the indoor enclosed space, m s Indicates the air supply coverage of the environmental control equipment, W s Indicates the absolute humidity of the supply air, m r Indicates the return air volume of the environmental control equipment, m in Indicates the amount of infiltrated air, W out represents the outdoor absolute humidity, and G represents the indoor humidity generation rate. This model needs to be combined with the psychological humidity chart to handle the conversion between relative humidity and absolute humidity.

[0171] The air quality model focuses on the changes in indoor CO2 concentration, because CO2 concentration is often used as an indicator of indoor air quality. This model needs to consider the CO2 produced by human breathing, the CO2 taken away by the ventilation system, and the CO2 brought by outdoor air infiltration. The basic equation of the air quality model can be expressed as:

[0172]

[0173] Where: C in Indicates the CO2 concentration in an indoor enclosed space, G CO2 represents the CO2 generation rate of the indoor enclosed space, Q represents the ventilation volume of the indoor enclosed space, C out represents the outdoor CO2 concentration. This model can be extended to handle other air pollutants such as particulate matter, volatile organic compounds, etc.

[0174] The lighting model describes the distribution and variation of indoor illumination levels, and needs to consider the combined effects of natural light and artificial lighting. A basic lighting model can be expressed as:

[0175] E=k1L+k2I s +δ

[0176] Where: E represents the indoor illumination of the indoor enclosed space, L represents the brightness of the artificial lighting controlled by the environmental control equipment, I s represents outdoor solar irradiance, k1 and k2 represent the equipment lighting influence coefficient and natural lighting influence coefficient respectively, and δ represents the adjustment coefficient. The equipment lighting influence coefficient, natural lighting influence coefficient and adjustment coefficient need to consider the direction and size of the window and the influence of shading facilities.

[0177] The parameters of these physical models (such as thermal conductivity, humidity generation rate, etc.) need to be determined through experimental measurements or data fitting. These parameters can be estimated based on historical data using optimization algorithms such as the least squares method. In the modeling process, the interactions between different models also need to be considered. For example, temperature changes affect relative humidity, humidity changes affect thermal comfort, C02 concentration affects ventilation strategies, etc. These interactions can be handled by coupled equations or iterative solutions. The advantage of physical models is that they are highly interpretable, can reflect the inherent mechanism of the system, and can still provide a reasonable description when data is scarce. For example, even in a new building without historical data, a preliminary physical model can be constructed as long as the physical properties of the building materials and the spatial geometry are known. In addition, physical models can also be used to simulate extreme situations or rare events, which are difficult to achieve in purely data-driven methods.

[0178] Next, the gray-box model method is used to fuse the data association-driven model and the indoor space physical model into an indoor comprehensive environmental model of the indoor enclosed space. The gray-box model is a modeling method that combines the advantages of the white-box model (a model based on physical principles) and the black-box model (a purely data-driven model). This fusion method aims to make full use of the theoretical basis of the physical model and the learning ability of the data-driven model, while achieving the effect of saving computing power and improving computing efficiency. The core idea of ​​the gray-box model is to use a data-driven approach to compensate or correct the uncertainties and simplified assumptions in the physical model. The specific implementation can use the data-driven model to predict the residuals (i.e., prediction errors) of the physical model. This approach allows the physical model to capture the main trends, while the data-driven model is responsible for subtle adjustments and unmodeled effects.

[0179] The advantage of this gray-box model approach is that it can more accurately describe the dynamics of complex indoor environments by combining the theoretical basis of physical models and the learning ability of data-driven models. The physical model provides an understanding of the basic behavior of the system, allowing the model to make reasonable predictions when faced with new situations. The introduction of the physical model reduces the reliance on large amounts of training data, especially in extreme or rare cases. The physical part of the model provides clear cause-effect relationships, which helps to understand and interpret the prediction results. Through reasonable structural design and optimization strategies, the gray-box model can significantly improve computational efficiency while maintaining high accuracy. In practical applications, this gray-box model can predict changes in indoor environmental parameters based on the current state and future control strategies, provide a basis for optimizing control, and detect system anomalies or equipment failures by comparing model predictions with actual measurements. It can also simulate energy consumption under different control strategies, find the best energy-saving solution, and evaluate and optimize indoor comfort by combining parameters such as temperature, humidity, and air quality.

[0180] In another implementation, a data-driven model with physical constraints can also be used for model fusion. Specifically, physical constraints are introduced in the training process of the data-driven model to ensure that its output conforms to basic physical laws. For example, a physical constraint term can be added to the loss function: L_total = L_data + λ*L_physics, where L_data is the loss based on data (such as mean square error), L_physics is the physical constraint loss (such as the degree of violation of the energy balance equation), and λ is the balance coefficient. In specific implementation, the following steps can be adopted: a) Initialization: Use the structure of the physical model as the initial architecture of the neural network, and initialize the network weights with the parameters of the physical model. This ensures that the model has a reasonable physical basis at the beginning of training. b) Data preparation: Combine historical equipment control parameters, indoor environmental data, and outdoor environmental data into a training data set. Ensure that the data set covers various operating conditions and seasonal changes. c) Model training: Use the prepared data set to train the gray box model. The training process can be divided into two stages: first, fix the physical model part and only train the data-driven part; then perform end-to-end fine-tuning, allowing the parameters of the physical model to be adjusted slightly. d) Online learning: Design an online learning mechanism that allows the model to be continuously updated and optimized during actual operation. This can be achieved by regularly collecting new data and performing incremental learning. e) Model evaluation: Use an independent test set to evaluate model performance. In addition to commonly used statistical indicators (such as RMSE, MAE), you should also check whether the model output conforms to physical laws, such as conservation of energy, conservation of mass, etc.

[0181] In one embodiment, the gray box model method is used to merge the data association driven model and the indoor space physical model into an indoor comprehensive environment model of the indoor closed space, including the following steps:

[0182] For any indoor space physical model, a new model error term is introduced into the indoor space physical model;

[0183] The data association driven model and the indoor space physical model are combined to obtain the initial indoor comprehensive environment model of the indoor enclosed space;

[0184] After preprocessing the historical equipment control parameters and the historical space environment data, the historical equipment control parameters and the historical space environment data are integrated into a model training set;

[0185] The model training set is input into the initial indoor comprehensive environment model, and the environment association change results are output through the data association driving model based on the historical equipment control parameters. The environment association change results and the historical space environment data are combined and the indoor space physical model is used to output the indoor space environment prediction data;

[0186] The mean square error between the indoor space environment prediction data and the historical indoor environment data is used as the model loss function;

[0187] Simultaneously adjusting the model error term of the indoor space physical model and the model parameters of the data association driven model according to the model loss function;

[0188] Continue to use the model training set to train the initial indoor comprehensive environment model until the model loss function reaches the minimum value, and obtain the trained indoor comprehensive environment model.

[0189] In this embodiment, for any indoor space physical model, the new model error term ε(t) is introduced to improve the flexibility and adaptability of the model. These error terms ε(t) can be designed as a function of time, for example, Fourier series expansion can be used to represent periodic changes:

[0190] ε(t)=a0+∑(a n *cos(nωt)+b n *sin(nωt))

[0191] where a0, a n , b n is a learnable parameter and ω is the fundamental frequency (e.g., a period of one day). This representation allows the error term to capture variations at different time scales.

[0192] The introduction of error terms increases the flexibility of the physical model, allowing the model to adapt to complex situations in the actual environment. Secondly, by learning these error terms, potential defects of the physical model can be discovered, providing guidance for further improvement of the physical model. Finally, this method retains the interpretability of the physical model while improving the accuracy of the model. In practical applications, by analyzing the learned error terms, unknown factors or neglected physical processes that affect the indoor environment can be identified, thereby continuously improving the understanding of the dynamics of the indoor environment.

[0193] Next, we combine the flexibility of data-driven with the theoretical basis of physical models, specifically creating a unified framework in which data-driven models and physical models can complement each other and work together. In specific implementation, the following methods can be used: First, the output of the data association-driven model (such as the CNN-LSTM network) is used as the input or parameter of the physical model. For example, the data-driven model can predict the dynamic changes of the thermal conductivity coefficient UA, and then input this predicted value into the thermal balance equation of the physical model. At the same time, the intermediate states and outputs of the physical model can be used as additional inputs to the data-driven model to form a closed-loop system.

[0194] This combination can be expressed as a set of coupled equations:

[0195] Y p =f p (X,θ p , Y NN )

[0196] Y NN =f NN (X,θ NN , Y p )

[0197] Y f =g(Y p , Y NN )

[0198] Among them, X is the input variable (such as equipment control parameters, environmental conditions, etc.), Y p is the output of the physical model, Y NN is the output of the neural network model, θ p and θ NN are the parameters of the physical model and the neural network model, respectively, p and f NNdenote the functions of the physical model and the neural network model, respectively, and g is the function that combines the outputs of the two models. A key advantage of this joint approach is that it allows the model to flexibly rely on physical principles or data-driven predictions in different situations. For example, in common situations with sufficient data, the model relies more on the data-driven part; while when the data is sparse or facing new situations, the model can rely more on physical principles. This adaptive ability greatly enhances the generalization performance and robustness of the model.

[0199] Next, you need to create a high-quality, information-rich dataset based on historical data so that it can fully reflect the dynamic characteristics of the indoor environment and the impact of control parameters. In specific implementation, you first need to identify and process outliers, missing values, and noisy data. For outliers, you can use statistical methods (such as Z-sCore or IQR methods) to detect them, and then decide whether to delete or replace them according to the specific situation. For missing values, you can use interpolation methods (such as linear interpolation or more complex time series interpolation methods) to fill them. Then make sure that all data are on the same time scale and that timestamps are aligned. This involves resampling or interpolating data with different sampling frequencies. Create new features or transform existing features to enhance the learning ability of the model. For example, you can add time-related features (such as time of day, day of the week, season, etc.), or create lagged features to capture time dependencies. Transform features of different scales to the same scale so that the model can learn the importance of each feature fairly. Divide the processed dataset into training, validation, and test sets. Usually, a ratio of 70% / 15% / 15% can be used, or a time-series split method can be used to simulate actual application scenarios. Perform appropriate augmentation on the training data to improve the generalization ability of the model. For example, small random perturbations can be added to simulate sensor noise:

[0200] The integrated model training set should contain the following components:

[0201] Input features: including pre-processed equipment control parameters (such as air conditioning set temperature, wind speed, etc.) and environmental parameters (such as outdoor temperature, humidity, sunlight intensity, etc.).

[0202] Target variables: pre-processed indoor environmental state parameters (such as indoor temperature, humidity, CO2 concentration, etc.).

[0203] Temporal information: used to capture time dependencies and periodic patterns.

[0204] Additional engineering features: hysteresis features, statistical features, etc.

[0205] The preprocessing and integration process improves data quality and eliminates noise and anomalies that affect model learning. Through feature engineering and selection, the expressiveness of data is enhanced, making it easier for the model to capture potential patterns and relationships. Standardization ensures that different features are treated fairly in the model. A reasonable data segmentation strategy provides a reliable foundation for model training, verification, and testing.

[0206] Next, the model training set is input into the initial indoor comprehensive environment model, and the calculation is performed through the data association driven model and the indoor space physical model. This step realizes the forward propagation of the model, generates prediction results, and lays the foundation for subsequent error calculation and parameter optimization. The specific implementation process is as follows:

[0207] The preprocessed model training set is input into the initial indoor comprehensive environment model in batches. Each batch contains a series of time step data, for example, 24 hours can be selected as a sequence length. The input data X includes historical equipment control parameters (such as air conditioning set temperature, wind speed, etc.) and historical space environment data (such as outdoor temperature, humidity, etc.). The historical equipment control parameters are processed by a data association driven model (such as a CNN-LSTM network) to output the environmental association change results. This process can be expressed as: NN =f NN (X c ,θ N N). Among them, X c is the device control parameter, Y NN is the predicted result of environmental correlation change. Then the result of environmental correlation change Y NN and historical spatial environment dataX enυ Combined as the input of the indoor space physical model. This combination process can be expressed as: p =g(Y NN , X enυ ). Among them, g is a combination function, which can be a simple splicing or a more complex fusion method. Then use the prepared input data to calculate the indoor space environment prediction data through the indoor space physical model. This process can be expressed as: p =f p (X p ,θ p , ε). Among them, θ p is the parameter of the physical model, and ε is the model error term. Finally, the output of the data association driven model and the output of the indoor space physical model are fused to obtain the final indoor space environment prediction data. The fusion method can be a simple weighted average or a more complex nonlinear combination.

[0208] For each time step and each predicted variable (such as temperature, humidity, C02 concentration, etc.), calculate the difference between the predicted value and the actual observed value. Assume that Y pred is the model prediction value, Y true is the actual observed value, then the error can be expressed as: e = Y pred -Y trye . The errors of all time steps and all predicted variables are squared and then averaged to get the mean square error between the predicted value and the actual value. To prevent overfitting, a regularization term can be added to the loss function. To ensure that the model prediction results conform to the laws of physics, a physical constraint term can be added to the loss function. Taking all the above factors into consideration, the final loss function can be constructed.

[0209] Next, we need to calculate the gradient of the loss function L with respect to each parameter. This includes the gradient of the physical model error term ε and the data association driven model parameter θ NN The gradient of . Using the chain rule, we get:

[0210]

[0211] Then the gradient descent method is used to update the parameters. For the physical model error term and the data-driven model parameters, the update rules are as follows:

[0212]

[0213] Among them, η ε and η NN is the learning rate, which can be fixed or adaptive. are the updated physical model error term and data-driven model parameters, respectively. In order to improve training efficiency and stability, an adaptive learning rate method, such as the Adam optimizer, can be used. When updating the physical model error term ε, it is necessary to ensure that the updated physical model still satisfies the basic laws of physics.

[0214] The model training set is continuously used to train the initial indoor comprehensive environment model. By adjusting the physical model error term and the data-driven model parameters at the same time, the coordinated optimization of the physical model and the data-driven model is achieved. When the model loss function reaches the minimum value, the trained indoor comprehensive environment model is obtained. This method not only retains the interpretability of the physical model, but also utilizes the flexibility and learning ability of the data-driven method, greatly improving the efficiency of the overall model training.

[0215] In one embodiment, after preprocessing the historical equipment control parameters and the historical space environment data, integrating the historical equipment control parameters and the historical space environment data into a model training set includes the following steps:

[0216] Unify the data sampling rates of historical equipment control parameters and historical spatial environment data;

[0217] Align the data of historical equipment control parameters at each time node with the data of historical spatial environment data at the previous time node;

[0218] The historical equipment control parameters and historical spatial environment data after data alignment are integrated into the model training set.

[0219] In this embodiment, unifying the data sampling rate of historical equipment control parameters and historical spatial environment data is a key step in data preprocessing. When it is implemented specifically, a target sampling rate must first be determined. The selection of this sampling rate should take into account the characteristics of indoor environmental changes and the response time of the control system. Generally, for building environmental control, a sampling interval of 5 to 15 minutes is more appropriate. Assume that 10 minutes is selected as a unified sampling interval. For data with a sampling rate higher than the target sampling rate, it is necessary to perform downsampling processing by using methods such as the average method, the median method, and the nearest neighbor method. For data with a sampling rate lower than the target sampling rate, interpolation processing is required. Common interpolation methods include linear interpolation, spline interpolation, nearest neighbor interpolation, etc. In addition, it is also necessary to process possible missing data. For short-term missing data (such as a single time point), interpolation methods can be used to fill in. For long-term missing data, historical data from the same period or statistical models can be considered for estimation. This step ensures the consistency of all data on a time scale, which is crucial for subsequent data analysis and model training. The unified sampling rate allows different types of data to be directly compared and correlated.

[0220] Next, align the data of historical device control parameters at each time node with the data of historical spatial environment data at the previous time node. In specific implementation, a new data structure needs to be created, in which each data point contains the device control parameters at the current moment and the spatial environment data at the previous moment. Assume the following data sequence:

[0221]

[0222] The aligned data structure should be as follows:

[0223]

[0224] The specific steps to implement this alignment process are as follows:

[0225] Create a new data frame with the following columns: timestamp, all control parameters, all environmental parameters (with "_prev" suffix). For each time point t(i), copy all control parameter values ​​at time t(i) to the new data frame, copy all environmental parameter values ​​at time t(i-1) to the new data frame, add the "_prev" suffix, and set the timestamp to t(i). For the first time point, since there is no environmental data at the previous moment, the environmental data at the first time point can be used as the initial value. Data alignment establishes the temporal relationship between control operations and environmental changes, allowing the model to learn the impact of control operations on the environment. This alignment also takes into account the hysteresis of indoor environmental changes, which is more in line with actual physical processes. For example, after adjusting the set temperature of the air conditioner, the indoor temperature will not change immediately, but will only show the effect after a certain period of time.

[0226] Next, the historical equipment control parameters and historical spatial environment data after data alignment are integrated into the model training set. In specific implementation, it is necessary to first design a data structure suitable for machine learning model input. Usually, this structure can be a two-dimensional table, in which each row represents a time point, and the columns include timestamp, all equipment control parameters at the current moment (such as air conditioning set temperature, wind speed, fresh air volume, etc.), all spatial environment data at the previous moment (such as indoor temperature, humidity, CO2 concentration, etc.), outdoor environment data at the current moment (such as outdoor temperature, humidity, sunshine intensity, etc.), and other related features (such as time features, building features, etc.).

[0227] Then transform the features of different scales to the same scale so that the model can learn the importance of each feature fairly. Divide the integrated dataset into training set, validation set and test set. The commonly used division ratio is 60%:20%:20% or 70%:15%:15%. For time series data, time order division is usually adopted to simulate the actual prediction scenario. Convert the processed data into a format that can be directly used by machine learning frameworks (such as PyTorch, TensorFlow), usually in tensor format.

[0228] In one embodiment, taking the comprehensive objective function as the algorithm optimization target, generating the initial device control strategy of the environment control device based on the indoor comprehensive environment model and using the ant colony optimization algorithm includes the following steps:

[0229] Initialize the algorithm parameters of the ant colony optimization algorithm;

[0230] The ant positions of multiple ants are randomly generated in the state space based on the constraint conditions, and different ant positions represent different device control parameters;

[0231] For each ant position, the corresponding comprehensive objective function value is calculated using the indoor comprehensive environment model;

[0232] Update the pheromone concentration in the state space, and update the ant position based on the comprehensive objective function value and the updated pheromone concentration;

[0233] Repeat the above steps of calculating the comprehensive objective function value and updating the ant positions until the maximum number of iterations of the ant colony optimization algorithm is reached, and output the optimal ant path under the constraints;

[0234] The optimal parameter range of each device control parameter of the environmental control device is determined according to the optimal ant path, and the optimal parameter range of each device control parameter of the environmental control device is integrated into an initial device control strategy.

[0235] In this implementation, the following key parameters need to be initialized before the ant colony optimization algorithm is executed:

[0236] Number of ants (m): determines the number of ants that explore the solution space simultaneously in each iteration. It is usually set between 20-100, adjusted according to the size and complexity of the problem.

[0237] Maximum number of iterations (Max_Iter): One of the conditions for setting the algorithm to stop, usually set between 100-1000, depending on the complexity of the problem and the desired accuracy.

[0238] Pheromone importance factor (α): controls the influence of pheromones on ants' path selection, usually set between 1-5.

[0239] Heuristic factor importance (β): controls the influence of heuristic information on ants' path selection, usually set between 1-5.

[0240] Pheromone volatility coefficient (ρ): controls the volatility of pheromones and is usually set between 0.1 and 0.9. A smaller value will speed up the convergence, but may fall into a local optimum; a larger value will help the global search, but may slow down the convergence.

[0241] Initial pheromone concentration (τ0): Sets the initial pheromone concentration in the state space, usually set to a small positive number such as 0.1.

[0242] Global optimal solution update weight (Q): The amount of pheromone increase used to update the global optimal solution, usually set to 1.

[0243] The initialization of these parameters can be expressed as:

[0244] params={'m':50,'Max_Iter':500,'alpha':2,'beta':3,'rho':0.5,'tau0':0.1,'Q':1}

[0245] The specific choice of parameters needs to be adjusted according to the characteristics of the problem. For example, for a multi-dimensional optimization problem such as indoor environment control, a larger number of ants and a larger number of iterations may be required to ensure that the solution space is fully explored. In addition, problem-specific parameters need to be initialized, such as the upper and lower bounds of the control variables, constraints, etc. For example, for air conditioning temperature setting: temp_bounds = [18, 28] # temperature setting range (degrees Celsius).

[0246] Then, based on the constraints, the ant positions of multiple ants are randomly generated in the state space. Specifically, different ant positions represent different device control parameters. According to the characteristics of the problem, the dimension and range of the state space are determined. For indoor environmental control, the state space may include multiple dimensions such as temperature setting, humidity setting, wind speed, and fresh air volume. Consider the physical limitations of the equipment and environment, such as temperature setting range, humidity control capability, energy consumption limit, etc. These constraints will limit the feasible range of ant positions. When the constraints are met, a random number generator is used to generate an initial position for each ant.

[0247] The position of each ant (i.e., a set of specific device control parameters) is input into the pre-trained indoor comprehensive environment model, the corresponding environmental state and performance indicators are calculated, and then these indicators are integrated into a single evaluation value through the comprehensive objective function. Specifically, for each ant i, its position is represented by a set of control parameters X_i = [T_i, RH_i, V_i, Q_i, ...], and the calculation process is as follows:

[0248] Predict environmental status using comprehensive indoor environmental model:

[0249] [T_indoor,RH_indoor,CO2]=IndoorModel(X_i,Outdoor_Conditions)

[0250] Calculate various performance indicators:

[0251] PMV=f_PMV(T_indoor, RH_indoor, V_i,...)

[0252] PPD=f_PPD(PMV)

[0253] Energy=f_Energy(X_i, T_indoor, RH_indoor,...)

[0254] IAQ = f_IAQ(CO2)

[0255] Calculate the comprehensive objective function value:

[0256] O=w1*abs(PMV)+w2*PPD+w3*Energy+w4*IAQ

[0257] Among them, w1, w2, w3, and w4 are the weights of each indicator, which can be adjusted according to specific needs.

[0258] By using a comprehensive environmental model, the indoor environmental state under different control parameters can be accurately predicted, avoiding the cost and time consumption of actual experiments. This greatly improves the efficiency and feasibility of the optimization process. The use of a comprehensive objective function allows for trade-offs between multiple objectives such as comfort, energy consumption, and air quality. By adjusting the weights, it can be flexibly adapted to different optimization needs. This step provides the necessary basis for subsequent pheromone updates and ant position updates.

[0259] Next, it is necessary to update the pheromone concentration in the state space, and update the ant position in combination with the comprehensive objective function value and the updated pheromone concentration. Specifically, the pheromone concentration is updated according to the path quality of each ant (i.e., the comprehensive objective function value). The better the path quality, the more pheromones are added. During the update process, it is also necessary to simulate the evaporation of pheromones in nature to prevent the algorithm from converging to a local optimal solution too early. Finally, based on the updated pheromone concentration and heuristic information, a new ant position is generated. Repeat the above steps of calculating the comprehensive objective function value and updating the ant position, and continuously improve the quality of the solution until the maximum number of iterations of the ant colony optimization algorithm is reached, converge to a near-optimal solution, and output the optimal ant path under the constraints, so as to find the control parameter combination that can best balance comfort, energy consumption, and air quality.

[0260] Analyze the control parameter combination represented by the optimal ant path in detail. Evaluate the influence of each control parameter on the objective function. Define a suitable interval for each control parameter based on the optimal solution and sensitivity analysis results. Ensure that the parameter interval meets all necessary constraints. This ultimately realizes the conversion of the optimization algorithm results into a practical control parameter interval, achieving a bridge between theoretical optimization and practical application.

[0261] In one embodiment, the optimal parameter interval is used as the initial solution of the model predictive control strategy, the comprehensive objective function is used as the strategy optimization target of the model predictive control strategy, and the strategy prediction result is generated based on the indoor comprehensive environment model and through the model predictive control strategy, including the following steps:

[0262] The optimal parameter interval is used as the initial solution of the model predictive control strategy;

[0263] Based on the initial solution of the strategy and through the comprehensive indoor environment model, the prediction time domain and control time domain of the model predictive control strategy are determined;

[0264] On the basis of the prediction time domain, the comprehensive objective function is used as the strategy optimization target of the model predictive control strategy. The sequential quadratic programming method is used to rolling optimize the control sequence in the control time domain until the comprehensive objective function takes the minimum value to obtain the optimal control sequence. The optimal control sequence is used as the strategy prediction result generated by the model predictive control strategy.

[0265] In this embodiment, the optimal parameter interval is obtained from the initial device control strategy generated by the previous ant colony optimization algorithm. Specifically, for each control parameter of the environmental control device, there is a corresponding optimal parameter interval. For example, for air-conditioning temperature setting, the optimal parameter interval may be [22°C, 24°C]; for humidity control, it may be [40%, 60%]; for lighting brightness, it may be [300lux, 500lux]. Using these parameter intervals as the initial solution provides a near-optimal starting point for the model predictive control strategy. In the specific implementation process, it is also necessary to select the initial value from the initial solution of the strategy. The simplest method is to select the middle value of each parameter interval as the initial value. For multiple parameters, it can be expressed in vector form. In practical applications, weight factors can also be introduced to adjust the importance of different parameters.

[0266] Next, the prediction horizon and control horizon of the model predictive control strategy need to be determined based on the initial solution of the strategy and through the indoor comprehensive environment model. The prediction horizon refers to the length of time that the model predicts forward, while the control horizon refers to the length of time that the control input is actually optimized during this prediction process. The process of determining these two horizons first uses the indoor comprehensive environment model to simulate the response of the system under different control inputs. Based on the initial solution, multiple simulations are performed to observe the time required for the system to reach a steady state, which can be used as a reference for the prediction horizon. Generally, the prediction horizon should be long enough to capture the dynamic characteristics of the system, but not too long to avoid excessive computational burden. The control horizon is usually shorter than the prediction horizon and can be set to 1 / 3 to 1 / 2 of the prediction horizon. The choice of control horizon requires a balance between control flexibility and computational complexity. A longer control horizon can provide more optimization freedom, but it will also increase the computational burden; a shorter control horizon has a small amount of computation, but may result in less smooth control. In specific implementation, the following formula can be used: T p = k*τ, T c =m*τ. Where T p is the prediction time domain, T cis the control time domain, τ is the characteristic time constant of the system (which can be obtained through step response test), k and m are adjustment coefficients, usually k>m.

[0267] For indoor environmental control, τ may be in the range of 10-30 minutes, depending on factors such as room size and insulation performance. k can be selected between 3-5, and m can be selected between 1-2. For example, if τ = 20 minutes, k = 4, and m = 1.5, then T_p = 80 minutes and T_c = 30 minutes. In practical applications, adaptive methods can also be used to dynamically adjust the time domain. For example, the values ​​of k and m can be adjusted according to changes in system response speed and external disturbances. When the system changes slowly, the values ​​of k and m can be increased to reduce the calculation frequency; when the system changes drastically, the values ​​of k and m can be reduced to improve control accuracy.

[0268] Next, based on the prediction time domain, the comprehensive objective function is used as the strategy optimization target of the model predictive control strategy, and the control sequence in the control time domain is optimized by the sequential quadratic programming method. The sequential quadratic programming (SQP) method transforms the nonlinear optimization problem into a series of quadratic programming sub-problems. In each control cycle, the SQP algorithm first linearizes the system model and then solves the quadratic programming problem to obtain the optimal control sequence. This process is repeated until the convergence condition or the maximum number of iterations is reached. When implementing, the following iterative formula can be used: k +1=u k +α k *d k . Where u k is the control sequence for the kth iteration, d k is the search direction, α k is the step size. After each iteration, the system state is updated and the optimization window is rolled forward by one step.

[0269] The specific process is as follows: using the initial solution as the starting point, linearizing the indoor comprehensive environment model near the current working point, constructing a quadratic programming problem based on the linearized model and the comprehensive objective function, and using the quadratic programming solver to solve the optimal control increment d k. Use the above iterative formula to update the control sequence. If the change in the objective function value is less than the preset threshold or the maximum number of iterations is reached, stop. Apply the first control action in the optimized control sequence to the system. Move the time window forward one step to obtain the new system state and realize rolling optimization. This rolling optimization method can adapt to the dynamic changes and external interference of the system and improve the robustness of the control. For example, if a sudden increase in outdoor temperature is detected during the optimization process, the algorithm can quickly adjust the control strategy in the next optimization cycle and increase the cooling output in advance to maintain indoor comfort. The optimal control sequence finally obtained is the output of the model predictive control strategy, which takes into account the current state and foresees possible changes in the future, thereby achieving more intelligent and efficient environmental control.

[0270] In one embodiment, updating the algorithm parameters of the ant colony optimization algorithm according to the strategy prediction result, and using the updated ant colony optimization algorithm to generate the optimal device control strategy of the environmental control device includes the following steps:

[0271] Update the pheromone distribution of the ant colony optimization algorithm based on the strategy prediction results;

[0272] Generate the target constraints of the ant colony optimization algorithm based on the optimal parameter interval;

[0273] Taking the comprehensive objective function as the algorithm optimization target, the optimal equipment control strategy of the environmental control equipment is generated by combining the state space and target constraints and using the ant colony optimization algorithm.

[0274] In this embodiment, pheromone plays an important role in the ant colony algorithm, which represents the attractiveness of the path. The process of updating the pheromone distribution is actually adjusting the importance of different paths in the search space. In specific implementation, the strategy prediction results are first mapped to the search space of the ant colony algorithm. This can be achieved by establishing a corresponding relationship between the prediction results and the ant colony search space. For example, if the prediction results show that a certain temperature range is better, then the pheromone concentration on the path corresponding to this temperature range is increased. Secondly, the pheromone update formula is used to adjust the pheromone concentration: τij(t+1)=(1-ρ)τij(t)+Δτij, where τij is the pheromone concentration on the path (i, j), ρ is the pheromone volatility coefficient (usually between 0.1 and 0.5), and Δτij is the pheromone increment. Δτij can be set according to the quality of the prediction results, for example, it can be set as a function of the improvement degree of the prediction results: Δτij=f(ΔJ), where ΔJ is the improvement amount of the objective function. Finally, in order to prevent the algorithm from converging to the local optimum too early, the upper and lower limits of the pheromone can be set: τmin≤τij≤τmax. Through the above process, the advanced knowledge of model predictive control can be integrated into the ant colony algorithm, accelerating the convergence of the algorithm and improving the search efficiency. At the same time, the algorithm's adaptive ability is also enhanced, enabling it to better cope with dynamically changing environmental control needs.

[0275] Next, the target constraint of the ant colony optimization algorithm is generated according to the optimal parameter interval. In the specific implementation, for each control parameter i, its constraint is defined as: Li≤xi≤Ui, where Li and Ui are the lower and upper limits of the optimal parameter interval corresponding to parameter i, respectively. Then these constraints are combined into a constraint vector: C=[L1,U1,L2,U2,...,Ln,Un]. In the implementation of the ant colony algorithm, every time a new solution is generated, it is necessary to check whether these constraints are met. A penalty function can be defined to deal with the situation where the constraints are not met: P(x)=∑max(0,Li-xi,xi-Ui), where x is the current solution. This penalty function is added to the objective function: J′(x)=J(x)+λP(x), where J(x) is the original objective function and λ is the penalty factor. By adjusting the size of λ, the strictness of the constraints can be controlled. Finally, the comprehensive objective function is used as the algorithm optimization target, and the state space and target constraints are combined and the ant colony optimization algorithm is used to generate the optimal device control strategy for the environmental control device. The execution process of the ant colony optimization algorithm in this step can refer to the execution of the ant colony optimization algorithm in the aforementioned implementation. The optimal device control strategy finally generated contains the optimal control parameters of all device control parameters within the optimal parameter range.

[0276] The present invention also discloses an indoor space environment control system based on an optimization algorithm, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the indoor space environment control method based on the optimization algorithm as described in any one of the above-mentioned embodiments is implemented.

[0277] Among them, the processor can adopt a central processing unit (CPU). Of course, according to actual usage, other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. can also be adopted. The general-purpose processor can adopt a microprocessor or any conventional processor, etc., and this application does not impose any restrictions on this.

[0278] Among them, the memory can be an internal storage unit of a computer device, such as a hard disk or memory of a computer device, or an external storage device of a computer device, such as a plug-in hard disk, a smart memory card (SMC), a secure digital card (SD) or a flash memory card (FC) equipped on the computer device, etc., and the memory can also be a combination of an internal storage unit and an external storage device of a computer device. The memory is used to store computer programs and other programs and data required by the computer device. The memory can also be used to temporarily store data that has been output or is to be output, and this application does not impose any restrictions on this.

[0279] The present invention also discloses an environment control device, which includes the indoor space environment control system, air conditioning unit, lighting unit, ventilation unit, etc. disclosed in the above embodiments. Figure 2 The environmental control unit (i.e., environmental control equipment) includes an environmental control controller (i.e., indoor space environmental control system). The environmental control equipment is connected to the distribution box (input power DC600V) and the mains (input power AC380V) through two interfaces, ST1X1 and ST1X2. The socket models used by these two interfaces are XC158 / 24F4ZTP40 and XC158 / 22FSZTP40 respectively. This dual power supply design may be to ensure the stable operation of the system and the reliability of power supply. The environmental control equipment realizes communication with the host computer through two CAN interfaces, ST1X3 and ST1X4 (socket model XC158 / 22F10KTP40).

[0280] In addition, the environmental control device is also equipped with a connection cable interface ST1X6 (socket model XCE18F10KTP40) and a one-way valve interface ST1X7 (socket model XCE14F4KTP40). The cable interface ST1X6 can be connected to different unit equipment. Taking the air-conditioning unit as an example, the cable interface ST1X6 can be connected to the fan equipment of the air-conditioning unit. The environmental control device also includes a series of sensor terminals (STW1 to STW1-5). The sensors connected through the sensor terminals can detect temperature, humidity, air pressure, air quality and other data.

[0281] A person skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of protection of the present application is limited to these examples. In line with the concept of the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of different aspects of one or more embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.

[0282] One or more embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the present application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of the present application should be included in the protection scope of the present application.

Claims

1. A method for controlling indoor space environment based on optimization algorithm, characterized in that: The steps include: Acquire historical device control parameters of the environment control device and historical space environment data corresponding to the historical device control parameters, wherein the historical space environment data includes historical indoor environment data inside the indoor closed space where the environment control device is deployed and historical outdoor environment data outside the indoor closed space; Building an indoor comprehensive environment model of the indoor enclosed space by combining the historical equipment control parameters and the historical space environment data; Constructing a comprehensive objective function according to the indoor environmental comfort of the indoor enclosed space and the equipment energy consumption index of the environmental control equipment; Creating a state space of an ant colony optimization algorithm based on the device control parameter list of the environmental control device, and generating constraint conditions of the ant colony optimization algorithm according to parameter limits of each of the device control parameters in the device control parameter list; Taking the comprehensive objective function as the algorithm optimization target, generating an initial device control strategy for the environmental control device based on the indoor comprehensive environment model and using the ant colony optimization algorithm, wherein the initial device control strategy includes an optimal parameter range for each of the device control parameters; Using the optimal parameter interval as the initial solution of the model predictive control strategy, using the comprehensive objective function as the strategy optimization target of the model predictive control strategy, and generating a strategy prediction result based on the indoor comprehensive environment model and through the model predictive control strategy; The algorithm parameters of the ant colony optimization algorithm are updated according to the strategy prediction result, and the updated ant colony optimization algorithm is used to generate the optimal device control strategy of the environmental control device.

2. The indoor space environment control method based on optimization algorithm according to claim 1 is characterized in that: The step of combining the historical equipment control parameters and the historical space environment data to construct the indoor comprehensive environment model of the indoor closed space comprises the following steps: Constructing a data association driving model between the historical equipment control parameters and the historical spatial environment data based on a convolutional neural network and a long short-term memory network; Building an indoor space physical model of the indoor enclosed space in combination with the historical indoor environment data and the historical outdoor environment data, wherein the indoor space physical model includes a thermodynamic model, a humidity model, an air quality model, and a lighting model; The data association driven model and the indoor space physical model are integrated into an indoor comprehensive environment model of the indoor enclosed space by using a gray box model method.

3. The indoor space environment control method based on optimization algorithm according to claim 2 is characterized in that: The formula of the thermodynamic model is as follows: Where: T in represents the indoor ambient temperature of the indoor enclosed space, C represents the heat capacity of the indoor air, Q H represents the heat provided by the environmental control equipment, Q s Indicates outdoor solar radiation heat, Q in represents the heat generated by the indoor heat source, UA represents the total heat transfer coefficient of the building where the indoor enclosed space is located, T out Indicates the outdoor ambient temperature; The formula of the humidity model is as follows: Where: W in represents the indoor absolute humidity of the indoor enclosed space, V represents the indoor volume of the indoor enclosed space, m s represents the air supply coverage of the environmental control equipment, W s Indicates the absolute humidity of the supply air, m r Indicates the return air volume of the environmental control equipment, m in Indicates the amount of infiltrated air, W out represents the outdoor absolute humidity, and G represents the indoor humidity generation rate; The formula of the air quality model is as follows: Where: C in represents the CO2 concentration in the indoor enclosed space, G CO2 represents the CO2 generation rate of the indoor enclosed space, Q represents the ventilation volume of the indoor enclosed space, C out Indicates outdoor CO2 concentration; The formula of the lighting model is as follows: E=k1L+k2I s +δ Where: E represents the indoor illumination of the indoor enclosed space, L represents the brightness of the artificial lighting controlled by the environmental control device, I s represents outdoor solar irradiance, k1 and k2 represent equipment lighting influence coefficient and natural lighting influence coefficient respectively, and δ represents the adjustment coefficient.

4. The indoor space environment control method based on optimization algorithm according to claim 3 is characterized in that: The step of fusing the data association driven model and the indoor space physical model into an indoor comprehensive environment model of the indoor enclosed space by using a gray box model method comprises the following steps: For any of the indoor space physical models, a new model error term is introduced into the indoor space physical model; Combining the data association driven model with the indoor space physical model to obtain an initial indoor comprehensive environment model of the indoor enclosed space; After preprocessing the historical equipment control parameters and the historical space environment data, the historical equipment control parameters and the historical space environment data are integrated into a model training set; Inputting the model training set into the initial indoor comprehensive environment model, outputting the environment association change result based on the historical equipment control parameters and through the data association driving model, combining the environment association change result and the historical space environment data and outputting the indoor space environment prediction data through the indoor space physical model; The mean square error between the indoor space environment prediction data and the historical indoor environment data is used as a model loss function; Simultaneously adjusting the model error term of the indoor space physical model and the model parameters of the data association driven model according to the model loss function; The initial indoor comprehensive environment model is continuously trained using the model training set until the model loss function reaches a minimum value, thereby obtaining a trained indoor comprehensive environment model.

5. The indoor space environment control method based on optimization algorithm according to claim 4 is characterized in that: After preprocessing the historical equipment control parameters and the historical space environment data, integrating the historical equipment control parameters and the historical space environment data into a model training set comprises the following steps: Unifying the data sampling rates of the historical equipment control parameters and the historical spatial environment data; Align the data of the historical equipment control parameters in each time node with the data of the historical spatial environment data in the previous time node; The historical equipment control parameters and the historical spatial environment data after data alignment are integrated into a model training set.

6. The indoor space environment control method based on optimization algorithm according to claim 1 is characterized in that: The method of using the comprehensive objective function as the algorithm optimization target, generating the initial device control strategy of the environmental control device based on the indoor comprehensive environment model and using the ant colony optimization algorithm comprises the following steps: Initializing algorithm parameters of the ant colony optimization algorithm; Randomly generate ant positions of a plurality of ants in the state space based on the constraint condition, where different ant positions represent different device control parameters; For each of the ant positions, the corresponding comprehensive objective function value is calculated using the indoor comprehensive environment model; Updating the pheromone concentration in the state space, and updating the ant position in combination with the comprehensive objective function value and the updated pheromone concentration; Repeat the above steps of calculating the comprehensive objective function value and updating the ant positions until the maximum number of iterations of the ant colony optimization algorithm is reached, and output the optimal ant path under the constraint conditions; The optimal parameter interval of each device control parameter of the environmental control device is determined according to the optimal ant path, and the optimal parameter interval of each device control parameter of the environmental control device is integrated into an initial device control strategy.

7. The indoor space environment control method based on optimization algorithm according to claim 6 is characterized in that: The method of using the optimal parameter interval as the initial solution of the model predictive control strategy, using the comprehensive objective function as the strategy optimization target of the model predictive control strategy, and generating a strategy prediction result based on the indoor comprehensive environment model and through the model predictive control strategy comprises the following steps: Using the optimal parameter interval as the strategy initial solution of the model predictive control strategy; Determining the prediction time domain and control time domain of the model prediction control strategy based on the initial solution of the strategy and through the indoor comprehensive environment model; On the basis of the prediction time domain, the comprehensive objective function is used as the strategy optimization target of the model predictive control strategy, and the control sequence in the control time domain is rollingly optimized using the sequential quadratic programming method until the comprehensive objective function takes the minimum value to obtain the optimal control sequence, and the optimal control sequence is used as the strategy prediction result generated by the model predictive control strategy.

8. The indoor space environment control method based on optimization algorithm according to claim 7 is characterized in that: The updating of the algorithm parameters of the ant colony optimization algorithm according to the strategy prediction result, and the use of the updated ant colony optimization algorithm to generate the optimal device control strategy of the environmental control device comprises the following steps: Updating the pheromone distribution of the ant colony optimization algorithm according to the prediction result of the strategy; Generating target constraints of the ant colony optimization algorithm according to the optimal parameter interval; The comprehensive objective function is used as the algorithm optimization target, the state space and the target constraint conditions are combined and the ant colony optimization algorithm is used to generate the optimal device control strategy of the environmental control device.

9. An indoor space environment control system based on an optimization algorithm, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the indoor space environment control method based on the optimization algorithm as described in any one of claims 1 to 8 is implemented.

10. An environmental control device, characterized in that: It comprises an indoor space environment control system based on the optimization algorithm as described in claim 9.

Images