Region Power Simulation Method, Device, Equipment and Medium Based on Mobile Trajectory

Abstract

The present disclosure relates to a method, apparatus, device and medium for regional power simulation based on movement trajectories. The method includes: obtaining movement trajectory data of different user types in the activity area, and obtaining the building type of the activity area; predicting the behavior state of the activity area based on the building type and the time information of the trajectory points in the movement trajectory data to determine the behavior states of different user types; predicting the electrical appliance usage behaviors of different user types in the activity area according to the behavior states of different user types and the external environment parameters corresponding to the trajectory points; and performing energy consumption simulation on the activity area based on the electrical appliance usage behaviors of different user types to obtain energy consumption simulation data of the activity area. It can be seen that the movement trajectory data packet contains various microscopic features and complex features. Therefore, when facing the diversity and complexity of individual behaviors in the activity area, it is possible to accurately perform regional power simulation, ultimately improving the energy conservation and emission reduction effects, energy efficiency and management effects.

Description

Technical Field

[0001] The present disclosure relates to the field of power simulation technology, and in particular, to a method, device, equipment and medium for regional power simulation based on mobile trajectories. Background Art

[0002] In order to meet the needs of energy conservation, emission reduction, optimizing energy efficiency and management, etc., it is necessary to use computer simulation technology to analyze and simulate the energy systems of activity areas (such as buildings like residential buildings, schools, shopping malls, etc.) to find the best energy usage plan for energy consumption calculation.

[0003] Currently, the power simulation methods for activity areas mainly include data-driven, simulation software, and bottom-up algorithms with multiple features. However, the existing methods only consider the macroscopic energy consumption data of the activity areas for energy consumption calculation. Therefore, for activity areas with the diversity and complexity of individual behaviors, regional power simulation cannot be accurately performed, thus unable to ensure the effects of energy conservation, emission reduction, energy efficiency and management. Summary of the Invention

[0004] In order to solve the above technical problems or at least partially solve the above technical problems, the present disclosure provides a method, device, equipment and medium for regional power simulation based on mobile trajectories.

[0005] In a first aspect, the present disclosure provides a method for regional power simulation based on mobile trajectories, the method comprising:

[0006] Obtaining mobile trajectory data of different user types in an activity area, and obtaining the building type of the activity area;

[0007] Based on the building type and the time information of the trajectory points in the mobile trajectory data, predicting the behavior states of the activity area to determine the behavior states of different user types;

[0008] Predicting the electrical appliance usage behaviors of different user types in the activity area according to the behavior states of different user types and the external environment parameters corresponding to the trajectory points;

[0009] Based on the electrical appliance usage behaviors of different user types, performing energy consumption simulation on the activity area to obtain energy consumption simulation data of the activity area.

[0010] In a second aspect, the present disclosure provides a device for regional power simulation based on mobile trajectories, the device comprising:

[0011] An obtaining module, configured to obtain mobile trajectory data of different user types in an activity area, and obtain the building type of the activity area;

[0012] The first prediction module is used to predict the behavior state of the activity area based on the building type and the time information of the trajectory points in the movement trajectory data, and determine the behavior states of different user types.

[0013] The second prediction module is used to predict the electrical appliance usage behaviors of different user types in the activity area according to the behavior states of different user types and the external environment parameters corresponding to the trajectory points.

[0014] The simulation module is used to perform energy consumption simulation on the activity area based on the electrical appliance usage behaviors of different user types, and obtain the energy consumption simulation data of the activity area.

[0015] In a third aspect, an embodiment of the present disclosure further provides a device, which includes:

[0016] One or more processors;

[0017] A storage device for storing one or more programs,

[0018] When the one or more programs are executed by the one or more processors, the one or more processors implement the method provided in the first aspect.

[0019] In a fourth aspect, an embodiment of the present disclosure further provides a computer-readable medium, on which a computer program is stored, and when the computer program is executed by a processor, the method provided in the first aspect is implemented.

[0020] The technical solution provided by the embodiment of the present disclosure has the following advantages compared with the prior art:

[0021] A method, device, equipment and medium for regional power simulation based on movement trajectory in the embodiment of the present disclosure obtain the movement trajectory data of different user types in the activity area and obtain the building type of the activity area; predict the behavior state of the activity area based on the building type and the time information of the trajectory points in the movement trajectory data, and determine the behavior states of different user types; predict the electrical appliance usage behaviors of different user types in the activity area according to the behavior states of different user types and the external environment parameters corresponding to the trajectory points; perform energy consumption simulation on the activity area based on the electrical appliance usage behaviors of different user types, and obtain the energy consumption simulation data of the activity area. It can be seen that the movement trajectory data packet contains a variety of microscopic features and complex features. By performing regional power simulation on the movement trajectory data and the building type, the microscopic behaviors of individuals can be flexibly and accurately mined. Therefore, when facing the activity area with the diversity and complexity of individual behaviors, the regional power simulation can be accurately performed, and finally the energy conservation and emission reduction effect, energy efficiency and management effect are improved. Description of the Drawings

[0022] The accompanying drawings herein are incorporated into and constitute a part of this specification, showing embodiments consistent with the present disclosure and, together with the specification, are used to explain the principles of the present disclosure.

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure or the prior art, the following will briefly introduce the accompanying drawings required for use in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0024] Figure 1 It is a schematic flowchart of a method for regional power simulation based on a moving trajectory provided by an embodiment of the present disclosure;

[0025] Figure 2 It is a schematic flowchart of S120 provided by an embodiment of the present disclosure;

[0026] Figure 3 It is a schematic flowchart of another method for regional power simulation based on a moving trajectory provided by an embodiment of the present disclosure;

[0027] Figure 4 It is a schematic structural diagram of a second prediction model based on MAPPO provided by an embodiment of the present disclosure;

[0028] Figure 5 It is a schematic flowchart of yet another method for regional power simulation based on a moving trajectory provided by an embodiment of the present disclosure;

[0029] Figure 6 It is a schematic structural diagram of a device for regional power simulation based on a moving trajectory provided by an embodiment of the present disclosure;

[0030] Figure 7 It is a schematic structural diagram of a device for regional power simulation based on a moving trajectory provided by an embodiment of the present disclosure. Detailed implementation manners

[0031] In order to be able to more clearly understand the above objects, features, and advantages of the present disclosure, the following will further describe the solutions of the present disclosure. It should be noted that, without conflict, the embodiments of the present disclosure and the features in the embodiments can be combined with each other.

[0032] Many specific details are set forth in the following description in order to fully understand the present disclosure, but the present disclosure can also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only a part of the embodiments of the present disclosure, rather than all the embodiments.

[0033] Current regional power simulation methods mainly include data-driven, simulation software, and bottom-up algorithms with multiple features. Specifically, the core of the data-driven regional power simulation method is to use an artificial neural network to process the change patterns of historical energy consumption data in the time series to obtain energy consumption simulation data for the active area under different environmental impacts; the core of the regional power simulation method based on simulation software is to simulate the annual operating energy consumption of the entire building centered on restoring physical environmental conditions, power system, and building structure performance to obtain energy consumption simulation data for the entire building in the active area; the core of the bottom-up algorithm with multiple features is to calculate the energy consumption of the entire building based on the impact of changes in the environment and building occupants on the energy consumption of electrical equipment to obtain energy consumption simulation data for the active area.

[0034] It can be seen that the above three regional power simulation methods only consider the macroscopic energy consumption data of the active area (such as environmental factors of the building, annual energy consumption of the building, changes in building occupants, etc.) for energy consumption calculation, without considering the diversity of individual behaviors and the complexity of building types in the active area. Therefore, for the active area with the diversity of individual behaviors and the complexity of building types, accurate regional power simulation cannot be carried out, thus unable to ensure the energy conservation and emission reduction effects, energy efficiency, and management effects.

[0035] To improve the accuracy of regional power simulation based on mobile trajectories, the embodiments of the present disclosure provide a regional power simulation method, device, and medium based on mobile trajectories.

[0036] Next, first in combination with Figures 1 to 5 a regional power simulation method based on mobile trajectories provided by the embodiments of the present disclosure will be described.

[0037] Figure 1 The flowchart of a regional power simulation method based on mobile trajectories provided by the embodiments of the present disclosure is shown.

[0038] In the embodiments of the present disclosure, Figure 1 the regional power simulation method shown can be executed by a regional power simulation device based on mobile trajectories. The regional power simulation device based on mobile trajectories can be an electronic device or a server. The electronic device can include but is not limited to fixed terminals such as smartphones, laptops, and desktop computers. The server can be a cloud server or a server cluster and other devices with storage and computing functions. The embodiments of the present disclosure will be explained in detail with the electronic device as the execution subject.

[0039] As Figure 1 shown, the regional power simulation method based on mobile trajectories can include the following steps.

[0040] S110. Obtain the movement trajectory data of different user types in the activity area, and obtain the building type of the activity area.

[0041] In this embodiment, when power simulation (i.e., energy consumption simulation) needs to be performed on the activity area, the electronic device obtains the movement trajectory data of different user types in the activity area from a terminal running the Global Positioning System (GPS), and obtains the building type of the activity area, so as to achieve regional power simulation considering the diversity of individual behaviors and the complexity of the building types in the activity area.

[0042] Among them, different user types can be understood as users of different age groups. Optionally, the user types include users in the age group of 1 - 16 years old, users in the age group of 16 - 25 years old, users in the age group of 26 - 35 years old, users in the age group of 36 - 45 years old, and users over 45 years old. It should be noted that the behavior states of different user types are different at different times, and the family situations and income situations corresponding to different user types are also different.

[0043] Among them, the activity area is the spatial area where users often move. The building type is the type label of the spatial area. For example, the building types include residential, school, shopping mall, etc.

[0044] Among them, the movement trajectory data can be understood as the geographical spatio - temporal data in the geographical space environment. The movement trajectory data can include the time information corresponding to the time scale and the spatial information corresponding to the spatial scale.

[0045] It can be understood that the movement trajectory data includes not only the user activity situation at the macro level, but also the user activity patterns and electricity consumption patterns at the micro level. By analyzing the movement trajectory data, the daily behavior patterns and preferences of different user types can be understood, and then the main activity areas or hot spots of individual users or a group of users can be identified. These areas are the gathering places for the life, work, entertainment, etc. of individual users or a group of users.

[0046] S120. Based on the building type and the time information of the trajectory points in the movement trajectory data, predict the behavior states of the activity area and determine the behavior states of different user types.

[0047] It can be understood that the movement trajectory data can be composed of a series of continuous or discontinuous trajectory points. The electronic device identifies the time information of the trajectory points from the movement trajectory data, and combines the building type and the time information of the trajectory points. First, it predicts the behavior states of different user types, and then based on the behavior states of different user types, it further predicts the electrical appliance usage behaviors of different user types in the activity area.

[0048] Among them, the behavioral states of different user types can be understood as the activity situations of users. Optionally, the behavioral states of different user types include, but are not limited to, learning state, working state, eating state, cooking state, sleeping state, walking state, running state, etc.

[0049] In some embodiments, the electronic device first obtains a neural network model for predicting behavioral states, and then uses the building type and the time information of the trajectory points in the movement trajectory data as the input data of the neural network model, so that the neural network model outputs the behavioral states of different user types. Specifically, the neural network model can be trained using the building type of the historical activity area and the time information of the trajectory points in the historical trajectory data of different user types in the historical activity area.

[0050] In some other embodiments, the electronic device determines the first time interval corresponding to the time information of the trajectory points. For each user type, if there is a corresponding behavioral state in one or more time intervals included in the activity area of the building type, the electronic device obtains a second time interval that includes the first time interval or is consistent with the first time interval from the multiple time intervals of each user type, and uses the behavioral state corresponding to the second time interval as the behavioral state of each user type, so as to obtain the behavioral states of different user types.

[0051] S130. Predict the electrical appliance usage behaviors of different user types in the activity area according to the behavioral states of different user types and the external environment parameters corresponding to the trajectory points.

[0052] In this embodiment, the electronic device first obtains the external environment parameters corresponding to the trajectory points, and then, for each user type, with the goal of optimizing the power consumption close to the real power consumption in the activity area, mines the behavioral states of each user type and the external environment parameters corresponding to the trajectory points for power load simulation, so as to obtain the electrical appliance usage behaviors of each user type, in order to predict the electrical appliance usage behaviors of different user types in the activity area.

[0053] In this embodiment, the method for determining the external environment parameters corresponding to the trajectory points includes, but is not limited to, the following methods: mapping the trajectory points from the geographical space environment to the map, and determining the longitude and latitude information of the trajectory points on the map; based on the longitude and latitude information of the trajectory points on the map, reading the external environment parameters corresponding to the trajectory points from the environmental prediction data containing multiple longitude and latitude information.

[0054] Among them, the geographical space environment can be understood as the space corresponding to the GPS system, and its corresponding coordinate system is the coordinate system corresponding to the GPS system. The coordinate system corresponding to the map is the map coordinate system.

[0055] Among them, the external environmental parameters corresponding to the trajectory points refer to the environmental information of the longitude and latitude where the trajectory points are located. Optionally, the external environmental parameters include, but are not limited to, temperature, humidity, and illuminance.

[0056] Specifically, obtain the mapping relationship between the geospatial environment and the map, that is, determine the mapping relationship between the coordinate system corresponding to the GPS system and the map coordinate system. Then, based on this mapping relationship, map the trajectory points from the geospatial environment to the map to determine the longitude and latitude information of the trajectory points on the map. Next, since different longitude and latitude information corresponds to different environmental prediction data, based on the longitude and latitude information of the stop points on the map, read the environmental prediction data corresponding to the trajectory points from the environmental prediction data containing multiple longitude and latitude information as the external environmental parameters corresponding to the trajectory points.

[0057] Among them, the electrical appliance usage behaviors of different user types refer to the behaviors of whether to start the electrical appliances. Optionally, the electrical appliance usage behaviors can be represented by different labels.

[0058] In addition, factors such as the behavior status of each user type, the external environmental parameters corresponding to the trajectory points, the family situation and income situation of each user type can also be combined to perform power load simulation, so as to improve the power load simulation accuracy by combining multiple fine-grained influencing factors.

[0059] S140. Based on the electrical appliance usage behaviors of different user types, perform energy consumption simulation on the activity area to obtain the energy consumption simulation data of the activity area.

[0060] It can be understood that since the electrical appliance usage behaviors of different user types consider the individual micro-behaviors, after performing power simulation on the activity area based on the electrical appliance usage behaviors of different user types, accurate energy consumption simulation data can be obtained.

[0061] In this embodiment, the specific implementation manner of S140 includes, but is not limited to, the following manner: for each user type, multiply the electrical appliance usage behavior of each user type by the corresponding rated power to obtain the energy consumption simulation data of each user type; based on the energy consumption simulation data of each user type and the proportion of each user type among different user types, perform weighted summation calculation to obtain the energy consumption simulation data of the activity area.

[0062] Among them, the determination method of the proportion of each user type among different user types includes: determining the number of users of each user type in the activity area and the total number of users of different user types; dividing the number of users of each user type in the activity area by the total number of users of different user types to obtain the proportion of each user type among different user types.

[0063] Exemplarily, assume that the activity area includes three types of users, namely user type A, user type B, and user type C. The proportion of user type A among different user types is k1, the proportion of user type B among different user types is k2, and the proportion of user type C among different user types is k3. Among them, the label of the electrical appliance usage behavior of user type A is x, and the rated power corresponding to the electrical appliance usage behavior of user type A is P1. The label of the electrical appliance usage behavior of user type B is y, and the rated power corresponding to the electrical appliance usage behavior of user type B is P2. The label of the electrical appliance usage behavior of user type C is z, and the rated power corresponding to the electrical appliance usage behavior of user type C is P3. Then, the energy consumption simulation data of user type A = x * P1, the energy consumption simulation data of user type B = y * P2, and the energy consumption simulation data of user type C = z * P3; then, the energy consumption simulation data of the activity area = x * P1 * k1 + y * P2 * k2 + z * P3 * k3.

[0064] In this way, by combining the electrical appliance usage behavior of each user type, the rated power corresponding to the electrical appliance usage behavior of each user type, and the proportion of each user type among different user types, power simulation is performed on the activity area. This power simulation method is simple and accurate and can be widely applied in the regional power simulation scenario.

[0065] A regional power simulation method based on mobile trajectories according to an embodiment of the present disclosure obtains mobile trajectory data of different user types in an activity area and obtains the building type of the activity area; predicts the behavior state of the activity area based on the building type and the time information of the trajectory points in the mobile trajectory data to determine the behavior states of different user types; predicts the electrical appliance usage behavior of different user types in the activity area according to the behavior states of different user types and the external environment parameters corresponding to the trajectory points; and performs energy consumption simulation on the activity area based on the electrical appliance usage behavior of different user types to obtain the energy consumption simulation data of the activity area. It can be seen that the mobile trajectory data packet contains a variety of microscopic features and complex features. By performing regional power simulation on the mobile trajectory data and the building type, the microscopic behavior of individuals can be flexibly and accurately mined. Therefore, when facing an activity area with the diversity and complexity of individual behaviors, regional power simulation can be accurately performed, ultimately improving the energy conservation and emission reduction effect, energy efficiency, and management effect.

[0066] In another implementation manner of the present disclosure, before determining the behavior states of different user types, the mobile trajectory data can also be preprocessed so that the behavior states of different user types and the electrical appliance usage behavior of different user types are predicted based on the preprocessed mobile trajectory data, thereby improving the prediction efficiency of the regional power simulation as a whole.

[0067] In some embodiments of the present disclosure, since there are a large number of redundant trajectory points in the movement trajectory data, stop points can be identified from the trajectory points included in the movement trajectory data to avoid affecting the prediction efficiency of the regional power simulation due to the redundant trajectory points. Before executing S120, the method further includes: obtaining stop points from the trajectory points according to the time information and spatial information of the trajectory points; correspondingly, the specific implementation manner of S120 includes: predicting the behavior state of the activity area based on the time information and building type of the stop points in the movement trajectory data to determine the behavior states of different user types; correspondingly, the specific implementation manner of S130 includes: predicting the electrical appliance usage behaviors of different user types in the activity area according to the behavior states of different user types and the external environmental parameters corresponding to the stop points.

[0068] In some embodiments, an algorithm based on DBSCAN clustering is used to mine stop points in the movement trajectory data. The algorithm is specifically TrajDBSCAN (Trajectory Density-Based Spatial Clustering of Applications with Noise). Specifically, the movement trajectory data further includes the spatial information of the trajectory points; then, obtaining stop points from the trajectory points according to the time information and spatial information of the trajectory points is specifically determined by the following method: for any current trajectory point in the movement trajectory data, obtaining the neighboring trajectory points of the current trajectory point according to the spatial information of the current trajectory point and the spatial information of other trajectory points; if the number of neighboring trajectory points is greater than or equal to the first quantity threshold, taking the current trajectory point as the current core point; if it is determined based on the time information of the current core point that the stay time of the current core point within a preset time period is greater than the preset time threshold, taking the current core point as the current stop point; updating the current trajectory point to the next trajectory point, and determining the next stop point based on the time information and spatial information of the next trajectory point until all stop points are obtained from the trajectory points.

[0069] Among them, the first quantity threshold is the minimum number of neighboring areas for determining core points from the trajectory points and can be determined according to experience.

[0070] Among them, a stop point refers to a core point indicating that an object stays at a certain position and maintains for a period of time. The preset time threshold is the minimum time for determining a stop point from the core points.

[0071] In some other embodiments, a stop point recognition method based on geometric characteristics and a stop point recognition method based on speed can also be used to obtain stop points from the trajectory points.

[0072] In this way, identifying stop points from the trajectory points included in the movement trajectory data helps to deeply understand the movement trajectory data. Therefore, the prediction efficiency and accuracy of the behavior states of different user types and the electrical appliance usage behaviors of different user types are improved, thereby improving the overall prediction efficiency of regional power simulation.

[0073] In this embodiment, after obtaining the stop points from the trajectory points, the method further includes: mapping the stop points from the geospatial environment to the map, and determining the longitude and latitude information of the stop points on the map; based on the longitude and latitude information of the stop points on the map, reading the external environment parameters corresponding to the stop points from the environment prediction data containing multiple longitude and latitude information.

[0074] Specifically, first, obtain the mapping relationship between the coordinate system corresponding to the GPS system and the map coordinate system, and then, based on this mapping relationship, map the stop points from the geospatial environment to the map to determine the longitude and latitude information of the stop points on the map. Then, since different longitude and latitude information corresponds to different environment prediction data, based on the longitude and latitude information of the stop points on the map, read the environment prediction data corresponding to the stop points from the environment prediction data containing multiple longitude and latitude information as the external environment parameters corresponding to the stop points.

[0075] Among them, the environment prediction data includes but is not limited to parameters such as temperature, humidity, and irradiance.

[0076] In this way, after determining the stop points, the external environment parameters corresponding to the stop points can be accurately determined based on the longitude and latitude information of the stop points on the map.

[0077] In the embodiment of the present disclosure, after determining the longitude and latitude information of the stop points on the map, the reverse parsing module of the map can also be used to determine the surrounding facilities of the stop points, such as catering, transportation, entertainment, etc., to provide more comprehensive information for different user types, which is convenient for better managing and using the stop points for behavior state prediction.

[0078] In some other embodiments of the present disclosure, when there is a positioning drift in the terminal running the GPS, there are noise data in the movement trajectory data collected by the terminal, and these noise data will affect the prediction accuracy of the behavior states of different user types.

[0079] To improve the prediction accuracy of the behavioral states of different user types, before executing S120, the method further includes: clustering the trajectory points in the mobile trajectory data according to the distribution density of the trajectory points in the mobile trajectory data to obtain at least one clustering cluster; obtaining the trajectory points that do not belong to the at least one clustering cluster from the mobile trajectory data, and removing the trajectory points that do not belong to the at least one clustering cluster as noise points; correspondingly, the specific implementation manner of S120 includes: predicting the behavioral states of the activity areas based on the building type and the time information of the trajectory points after removing the noise points, and determining the behavioral states of different user types; correspondingly, the specific implementation manner of S130 includes: predicting the electrical appliance usage behaviors of different user types in the activity area according to the behavioral states of different user types and the external environment parameters corresponding to the trajectory points after removing the noise points.

[0080] Specifically, the electronic device first uses the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm. Given the ε-neighborhood and the unvisited core points, by searching all reachable points in the ε-neighborhood of the core points, the neighborhood of the core points is obtained, and the set of core points is determined based on the neighborhood of the core points and the core points, that is, the distribution density of the trajectory points in the mobile trajectory data is obtained; then, according to the distribution density of the trajectory points in the mobile trajectory data, the trajectory points in the mobile trajectory data are clustered to obtain at least one clustering cluster; then, the trajectory points that do not belong to the at least one clustering cluster are obtained from the mobile trajectory data, and the trajectory points that do not belong to the at least one clustering cluster are removed as noise points to obtain the trajectory points after removing the noise points, and then the behavioral states of the activity areas are predicted based on the building type and the time information of the trajectory points after removing the noise points, and the behavioral states of different user types are determined.

[0081] Optionally, the neighborhood of the core points can be determined in the following way:

[0082] N ε (p) = {q|diss(p,q) ≤ ε}

[0083] where, N ε (p) is the ε-neighborhood of the core point p, q is the point in the ε-neighborhood of the core point p, and diss(p,q) is the distance between p and q.

[0084] Optionally, the set of core points can be determined in the following way:

[0085] Core(p) = {q||N ε (p)| ≥ MinPts}

[0086] Among them, Core(p) is the set of core points p, and MinPts is the preset sample point.

[0087] Among them, the clustering clusters can be determined in the following manner:

[0088]

[0089] Among them, D(p, q) is the clustering cluster containing p and q.

[0090] Furthermore, the trajectory points after removing the noise points are mapped from the geospatial environment to the map, and the longitude and latitude information of the trajectory points after removing the noise points on the map is determined; based on the longitude and latitude information of the trajectory points after removing the noise points on the map, the external environment parameters corresponding to the trajectory points after removing the noise points are read from the environmental prediction data containing multiple longitude and latitude information, and then the electrical appliance usage behaviors of different user types in the activity area are predicted by using the external environment parameters and the behavior states of different user types.

[0091] Thus, by clustering the trajectory points in the mobile trajectory data, the noise points are removed from the mobile trajectory data, and the electrical appliance usage behaviors are predicted by combining the external environment parameters corresponding to the trajectory points after removing the noise points and the behavior states of different user types. Therefore, the influence of noise points on the prediction of electrical appliance usage behaviors is avoided, thereby improving the prediction accuracy of electrical appliance usage behaviors.

[0092] It should be noted that the order of the above two preprocessing methods is not limited. That is to say, in this embodiment, the noise points can be removed from the mobile trajectory data before determining the stay points, so as to use the stay points after removing the noise points for behavior state prediction and electrical appliance usage behavior prediction; in this embodiment, the stay points can also be determined before removing the noise points, so as to use the stay points after removing the noise points for behavior state prediction and electrical appliance usage behavior prediction.

[0093] In this way, by preprocessing the mobile trajectory data and predicting the behavior states of different user types and the electrical appliance usage behaviors of different user types based on the preprocessed mobile trajectory data, the overall prediction efficiency of regional power simulation is improved.

[0094] In another implementation manner of the present disclosure, the specific implementation manner of determining the behavior states of different user types is explained in detail.

[0095] Figure 2 The flowchart of S120 provided by the embodiment of the present disclosure is shown.

[0096] As Figure 2 shown, S120 specifically includes the following steps.

[0097] S210. Simulate the time information and building type of the trajectory points onto different nodes of the preset graph structure in the first prediction model to obtain the current graph structure.

[0098] In this embodiment, the electronic device obtains the first prediction model. The first prediction model has constructed a preset graph structure, which includes different nodes and the pointing relationships between the nodes. After inputting the time information and building type of the trajectory points into the first prediction model, the time information and building type of the trajectory points are simulated onto different nodes of the preset graph structure to update the information represented by different nodes, thereby obtaining the current graph structure.

[0099] Among them, the first prediction model includes but is not limited to the Bayesian neural network model, and the preset graph structure includes but is not limited to the Directed Acyclic Graph (DAG). This embodiment combines the accurate description ability of the directed acyclic graph for data structures and the advantages of the Bayesian neural network model in uncertainty modeling to accurately identify the population activity situation, that is, to determine the probability values of the behavior states of different user types.

[0100] To better establish the probability relationships between the behavior states of different user types and the time information and building type of the trajectory points, the time reference information, building reference type, and behavior reference state are respectively abstracted onto different nodes of the preset graph structure in the first prediction model. Specifically, the node representing the time reference information points to the node representing the behavior reference state, which means that the change in time information will affect the change in the behavior state and its related probabilities; similarly, the node representing the historical building type points to the node representing the behavior reference state, which means that the change in building type will affect the change in the behavior state and its related probabilities. Thus, the preset graph structure in the first prediction model includes nodes corresponding to the time reference information, building reference type, and behavior reference state respectively, as well as the pointing relationship between the time reference information and the behavior reference state, and at the same time includes the pointing relationship between the building reference type and the behavior reference state.

[0101] Furthermore, after simulating the time information and building type of the trajectory points onto different nodes of the preset graph structure in the first prediction model, the time information and building type represented by the relevant nodes on the preset graph structure change accordingly, forming the current graph structure.

[0102] S220. Calculate the probability values of the behavior states of different user types based on the time information and building type represented by different nodes in the current graph structure.

[0103] In this embodiment, the current graph structure includes a first node abstracted from time information, a second node abstracted from building types, a first pointing relationship abstracted from time information and behavior states, and a second pointing relationship abstracted from building types and behavior states.

[0104] In this embodiment, the specific implementation manner of S220 includes but is not limited to the following: reading from the current graph structure a first node representing time information, a second node representing building types, a first pointing relationship corresponding to the time information, and a second pointing relationship corresponding to the building types; calculating probability values for the first node, the second node, the first pointing relationship, and the second pointing relationship according to the Bayesian probability calculation method to determine the probability values of the behavior states of different user types.

[0105] Optionally, when calculating probabilities by combining a directed acyclic graph and a Bayesian neural network model, the probability values of the behavior states of different user types can be determined in the following manner:

[0106]

[0107] where n is the number of nodes in the current graph structure, x i is the input data of the directed acyclic graph in the Bayesian neural network model, and this input data specifically includes the time information and building types of different user types, and π i is the set of parent nodes of x i .

[0108] In this way, the method for identifying population activity situations by integrating a directed acyclic graph and a Bayesian neural network model achieves the effect of accurately calculating the probability values of the behavior states of different user types based on mobile trajectory data.

[0109] S230. For each user type, obtain the behavior state with the maximum probability value from the probability values as the behavior state corresponding to each user type, and obtain the behavior states of different user types.

[0110] In this embodiment, starting from the first user type, for the current user type, obtain the behavior state with the maximum probability value from the probability values as the behavior state corresponding to the current user type, update the current user type, and obtain the behavior state with the maximum probability value from the probability values as the behavior state corresponding to the next user type until the behavior states corresponding to each user type are determined, and then obtain the behavior states of different user types.

[0111] In this way, using the behavior state with the maximum probability value as the behavior state corresponding to each user type ensures the accuracy of determining the behavior states of different user types.

[0112] It is understandable that the difference between the Bayesian neural network model and other neural network models lies in that the Bayesian neural network model takes into account the uncertainty of parameters and does not rely on a large amount of data for training.

[0113] In this embodiment, for the Bayesian neural network model, methods such as stochastic gradient descent or variational inference are usually used for model training. Specifically, taking the time reference information and the building reference type as the input data of the Bayesian neural network model, and the behavior reference state as the output data of the Bayesian neural network model, during the process of training the Bayesian neural network model using the input data and the output data, calculate the relative entropy (Kullback-Leibler divergence, KL divergence) of the Bayesian neural network model, and then based on the KL divergence, use methods such as stochastic gradient descent or variational inference to perform iterative training on the Bayesian neural network model until the iterative cutoff condition is met, and obtain the trained Bayesian neural network model.

[0114] Optionally, the KL divergence can be determined in the following manner:

[0115]

[0116] where \(q(w|\theta)\) is the distribution of the weight parameters after giving the parameters of the given normal distribution; is the likelihood of the observed data after giving the network parameters; \(P(w)\) is the prior of the weights; is the mathematical expectation.

[0117] Thus, in the case of less training data, choosing the Bayesian neural network model as the first prediction model can ensure the prediction accuracy of the behavior state to a certain extent.

[0118] In another implementation manner of the present disclosure, a detailed explanation is given for the specific implementation manner of the electrical appliance usage behaviors of different user types in the predicted activity area.

[0119] Figure 3 shows a schematic flowchart of another method for regional power simulation based on mobile trajectories provided by an embodiment of the present disclosure

[0120] As Figure 3 shown, the method for regional power simulation based on mobile trajectories specifically includes the following steps.

[0121] S310. Obtain the mobile trajectory data of different user types in the activity area, and obtain the building type of the activity area.

[0122] S320. Based on the building type and the time information of the trajectory points in the mobile trajectory data, perform a prediction on the behavior state of the activity area to determine the behavior states of different user types.

[0123] Among them, the specific implementation manners of S310 to S320 can be referred to the descriptions of the above embodiments and will not be elaborated here.

[0124] S330: Using the pre-determined behavior restriction range as a constraint condition, and using the second prediction model to process the behavior states of different user types and the external environment parameters corresponding to the trajectory points, so as to obtain the electrical appliance usage behaviors of different user types in the activity area.

[0125] Among them, the behavior restriction range can be understood as a constraint condition for narrowing the search range of electrical appliance usage behaviors. This behavior restriction range can exclude behaviors that do not conform to the rules or are impossible to lead to success, thereby reducing the search space.

[0126] Optionally, the behavior restriction range is the interval data selected from the fitting result after fitting the historical behaviors of the electrical appliances.

[0127] Among them, the second prediction model includes but is not limited to the model determined based on the multi-agent proximal policy optimization (MAPPO) algorithm. Specifically, each user type serves as an agent in MAPPO, and each agent maintains an action decision network, a behavior evaluation network, and an experience pool. The action decision network of each agent takes the behavior restriction range as a constraint condition and outputs an action according to the local observation of each agent to obtain a decision, that is, to obtain the operation strategy of each agent; the behavior evaluation network of each agent fits the global state value function to evaluate the environmental state, that is, to obtain the evaluation value corresponding to the operation strategy of each agent.

[0128] Then, in this embodiment, the specific implementation manner of S330 includes but is not limited to the following manner: Based on the action decision network in the second prediction model, process the behavior states of different user types and the external environment parameters corresponding to the trajectory points to obtain multiple operation behaviors within the behavior restriction range; Based on the behavior evaluation network in the second prediction model, calculate the evaluation values corresponding to the multiple operation behaviors respectively; According to the evaluation values corresponding to the multiple operation behaviors respectively, obtain the operation behavior with the highest evaluation value from the multiple operation behaviors as the electrical appliance usage behavior of different user types in the activity area.

[0129] For ease of understanding, refer to Figure 4 the structural schematic diagram of the second prediction model based on MAPPO shown in Figure 4The action decision-making network in it can be a combination of the action decision-making networks of multiple agents, and the evaluation network can be a combination of the evaluation networks of multiple agents. Specifically, input the behavior states of different user types and the external environmental parameters corresponding to the trajectory points into the second prediction model based on MAPPO. Then, based on the action decision-making network in the second prediction model, process the behavior states of different user types and the external environmental parameters corresponding to the trajectory points to obtain multiple operation behaviors within the behavior limit range. Then, based on the evaluation network in the second prediction model, calculate the evaluation values of the multiple operation behaviors to determine the evaluation values corresponding to the multiple operation behaviors respectively. Finally, obtain the operation behavior with the highest evaluation value from the multiple operation behaviors, so as to obtain the electrical appliance usage behaviors of different user types in the activity area.

[0130] In addition, factors such as the family situation and income situation of each user type can also be input into the above-mentioned second prediction model. Then, based on the action decision-making network in the second prediction model, process the behavior state, external environmental parameters, family situation, and income situation to obtain multiple operation behaviors within the behavior limit range, so as to improve the prediction accuracy of operation behaviors by combining various fine-grained influencing factors.

[0131] Thus, using different user types as the agents of the second prediction model to drive the action decision-making network and the behavior evaluation network, and using the behavior limit range as a constraint condition to process the behavior states of different user types and the external environmental parameters corresponding to the trajectory points, so as to accurately calculate the evaluation values corresponding to multiple operation behaviors respectively, improving the acquisition efficiency and accuracy of electrical appliance usage behaviors.

[0132] In this embodiment, after determining the electrical appliance usage behaviors of different user types, it can also include using the behavior states of different user types, the external environmental parameters corresponding to the trajectory points, and the electrical appliance usage behaviors of different user types as replay samples to optimize the second prediction model using the replay samples. Correspondingly, the method further includes:

[0133] S331. Take the behavior states of different user types, the external environment parameters corresponding to the trajectory points, and the electrical appliance usage behaviors of different user types as replay samples, and add the replay samples to the experience pool of the second prediction model; S332. Calculate the priorities of multiple replay samples to be sampled in the experience pool to determine the first priorities corresponding to the multiple replay samples to be sampled; S333. Sample multiple replay samples from the experience pool according to the first priorities corresponding to the multiple replay samples to be sampled; S334. Based on the behavior states and electrical appliance usage behaviors included in the multiple sampled replay samples, calculate the first loss data of the action decision network and the second loss data of the behavior evaluation network; S335. According to the first loss data and the second loss data, iteratively update the network parameters corresponding to the action decision network and the behavior evaluation network respectively to obtain the updated second prediction model.

[0134] In S331, the specific implementation method of adding the replay samples to the experience pool of the second prediction model includes, but is not limited to, the following methods: use the state value function calculation method to calculate the second priority of the replay samples; based on the second priority of the replay samples, add the replay samples to the experience pool of the second prediction model according to the data tree structure.

[0135] Among them, the state value function calculation method can be the priority calculation method of the action decision network, and this method can be expressed as Specifically, based on the second priority of the replay samples, abstract the replay samples into nodes of the data tree structure to realize adding the replay samples to the experience pool of the second prediction model according to the data tree structure.

[0136] Optionally, the data tree structure includes, but is not limited to, the binary tree structure (sumtree), and can also be other tree structures.

[0137] In S332, optionally, the first priorities corresponding to the multiple replay samples to be sampled can be determined in the following way:

[0138]

[0139] Among them, p(j) is the priority of the replay sample j to be sampled, p j is the corrected priority of the replay sample j to be sampled, p k is the priority of the replay sample k to be sampled, k is a random number, and α is the priority index.

[0140] Optionally, the corrected priority p j of the replay sample j to be sampled can be determined in the following way:

[0141] p j = |TD error (j)| + ε

[0142] Among them, TD error (j) is the temporal difference error of the replay sample j to be sampled, and ε is the relative error. Specifically, the larger TD error (j) is, the greater the room for improvement in prediction accuracy, the greater the role in backpropagation, the more useful information the algorithm can obtain from it, the higher the importance of the replay samples being learned, and these replay samples will be preferentially trained. The weights corresponding to these replay samples during training can be determined as follows:

[0143]

[0144] Among them, N is the size of the experience pool, representing the number of replay samples, and β represents the non-uniform probability compensation coefficient.

[0145] In S333, since the replay samples in the experience pool are stored according to the data tree structure, and each node of the data tree structure represents a replay sample with a priority. After determining the first priorities corresponding to multiple replay samples to be sampled, multiple replay samples are sampled from the experience pool based on the first priorities, and the sampled multiple replay samples are used as high-quality samples. Optionally, these high-quality samples can be stored in Figure 4 the high-quality experience pool of the experience pool shown.

[0146] In S334, optionally, the first loss data of the action decision network can be determined as follows:

[0147]

[0148]

[0149] Among them, L actor (θ i ) is the first loss data, is the mathematical expectation, is the behavior reference state, is the behavior state included in multiple replay samples, a i (t) is the action of the intelligent agent at time t, o i (t) is the observation information of the intelligent agent at time t, clip(*) is the pruning function, A i (t) is the use of generalized advantage estimation, ψ is the hyperparameter for balancing exploration and exploitation, ε is the relative error, S(o i (t)) is the action entropy at the observation o i (t).

[0150] In S334, optionally, the second loss data of the behavior evaluation network can be determined as follows:

[0151]

[0152] Among them, L critic (ξ i ) is the second loss data, s(t) is the entropy of the policy, is the calculation method of the state value function, and V i (*) is the state value function of the previous policy.

[0153] In S335, after determining the first loss data and the second loss data, the gradient descent parameters are determined based on the first loss data and the second loss data, and then the network parameters corresponding to the action decision network and the behavior evaluation network are iteratively updated using the gradient descent parameters to obtain the updated second prediction model.

[0154] In the above manner, after determining the electrical appliance usage behaviors of different user types in the activity area, the behavior states of different user types, the external environment parameters corresponding to the trajectory points, and the electrical appliance usage behaviors of different user types are used as replay samples to optimize the second prediction model. The optimized second prediction model has better performance than before, further improving the prediction accuracy of the electrical appliance usage behaviors of different user types.

[0155] S340. Based on the electrical appliance usage behaviors of different user types, perform energy consumption simulation on the activity area to obtain the energy consumption simulation data of the activity area.

[0156] Among them, the specific implementation manner of S340 can refer to the description of the above embodiments and will not be elaborated here.

[0157] In another implementation manner of the present disclosure, a specific explanation is given to one implementation manner of the regional power simulation method based on the movement trajectory.

[0158] Figure 5 Fig. shows a schematic flowchart of another regional power simulation method based on the movement trajectory provided by the embodiments of the present disclosure.

[0159] As Figure 5 shown, the regional power simulation method based on the movement trajectory specifically includes the following steps.

[0160] S510. Obtain the movement trajectory data of different user types in the activity area and obtain the building type of the activity area.

[0161] Among them, the movement trajectory data includes the time information and spatial information of the trajectory points.

[0162] S520. Perform noise removal processing on the movement trajectory data to obtain the trajectory points with noise points removed.

[0163] In this embodiment, the specific implementation manner of S520 includes but is not limited to the following:

[0164] Cluster the trajectory points in the movement trajectory data according to the distribution density of the trajectory points in the movement trajectory data to obtain at least one cluster;

[0165] Obtain the trajectory points that do not belong to at least one cluster from the movement trajectory data, and remove the trajectory points that do not belong to at least one cluster as noise points to obtain the trajectory points with noise points removed.

[0166] S530. Obtain the stop points from the trajectory points with noise points removed.

[0167] In this embodiment, the specific implementation manner of S530 includes but is not limited to the following: Obtain the stop points from the trajectory points with noise points removed according to the time information and spatial information of the trajectory points with noise points removed.

[0168] S540. Predict the behavior states of the activity areas based on the building type and the time information of the stop points in the movement trajectory data, and determine the behavior states of different user types.

[0169] In this embodiment, the specific implementation manner of S540 includes but is not limited to the following: Simulate the time information and building type of the trajectory points onto different nodes of a preset graph structure in a first prediction model to obtain a current graph structure; Calculate the probability values of the behavior states of different user types based on the time information and building type represented by different nodes in the current graph structure; For each user type, obtain the behavior state with the largest probability value from the probability values as the behavior state corresponding to each user type to obtain the behavior states of different user types.

[0170] S550. Predict the electrical appliance usage behaviors of different user types in the activity area according to the behavior states of different user types and the external environment parameters corresponding to the trajectory points.

[0171] In this embodiment, the specific implementation manner of S550 includes but is not limited to the following:

[0172] Using a second prediction model with a pre-determined behavior restriction range as a constraint condition, process the behavior states of different user types and the external environment parameters corresponding to the trajectory points to obtain the electrical appliance usage behaviors of different user types in the activity area.

[0173] Specifically, based on the action decision-making network in the second prediction model, the behavior states of different user types and the external environment parameters corresponding to the trajectory points are processed to obtain multiple operation behaviors within the behavior restriction range; based on the behavior evaluation network in the second prediction model, the evaluation values corresponding to the multiple operation behaviors are calculated; according to the evaluation values corresponding to the multiple operation behaviors, the operation behavior with the highest evaluation value is obtained from the multiple operation behaviors as the electrical appliance usage behaviors of different user types in the activity area.

[0174] S560. Based on the electrical appliance usage behaviors of different user types, perform energy consumption simulation on the activity area to obtain the energy consumption simulation data of the activity area.

[0175] In this embodiment, the specific implementation manner of S560 includes but is not limited to the following manner:

[0176] For each user type, multiply the electrical appliance usage behavior of each user type by the corresponding rated power to obtain the energy consumption simulation data of each user type;

[0177] Based on the energy consumption simulation data of each user type and the proportion of each user type among the different user types, perform weighted summation calculation to obtain the energy consumption simulation data of the activity area.

[0178] Through the above method, since the mobile trajectory data packet has various microscopic features and complex features, by performing regional power simulation on the mobile trajectory data and the building type, the microscopic behaviors of individuals can be flexibly and accurately mined. Therefore, when facing the diversity and complexity of individual behaviors in the activity area, accurate regional power simulation can be performed, ultimately improving the energy conservation and emission reduction effect, energy efficiency, and management effect. In addition, by preprocessing the mobile trajectory data and predicting the behavior states of different user types and the electrical appliance usage behaviors of different user types based on the preprocessed mobile trajectory data, the overall prediction efficiency of the regional power simulation is improved.

[0179] The embodiment of the present disclosure also provides a mobile-trajectory-based regional power simulation device for implementing the above-mentioned mobile-trajectory-based regional power simulation method. The following will be described in conjunction with Figure 6 For illustration. In the embodiment of the present disclosure, the mobile-trajectory-based regional power simulation device may be an electronic device. Among them, the electronic device may include devices with communication functions such as mobile terminals and tablet computers.

[0180] Figure 6 FIG. shows a structural schematic diagram of a mobile-trajectory-based regional power simulation device provided by an embodiment of the present disclosure.

[0181] As Figure 6 shown, the mobile-trajectory-based regional power simulation device 600 may include:

[0182] An acquisition module 610, configured to acquire the movement trajectory data of different user types in the activity area and acquire the building type of the activity area;

[0183] A first prediction module 620, configured to predict the behavior state of the activity area based on the building type and the time information of the trajectory points in the movement trajectory data, and determine the behavior states of different user types;

[0184] A second prediction module 630, configured to predict the electrical appliance usage behaviors of different user types in the activity area according to the behavior states of different user types and the external environment parameters corresponding to the trajectory points;

[0185] A simulation module 640, configured to perform energy consumption simulation on the activity area based on the electrical appliance usage behaviors of different user types, and obtain the energy consumption simulation data of the activity area.

[0186] A regional power simulation device based on movement trajectory according to an embodiment of the present disclosure acquires the movement trajectory data of different user types in the activity area and acquires the building type of the activity area; predicts the behavior state of the activity area based on the building type and the time information of the trajectory points in the movement trajectory data, and determines the behavior states of different user types; predicts the electrical appliance usage behaviors of different user types in the activity area according to the behavior states of different user types and the external environment parameters corresponding to the trajectory points; performs energy consumption simulation on the activity area based on the electrical appliance usage behaviors of different user types, and obtains the energy consumption simulation data of the activity area. It can be seen that the movement trajectory data packet contains various microscopic features and complex features. By performing regional power simulation based on movement trajectory on the movement trajectory data and the building type, the microscopic behaviors of individuals can be flexibly and accurately mined. Therefore, when facing an activity area with the diversity and complexity of individual behaviors, regional power simulation can be accurately performed, ultimately improving the energy conservation and emission reduction effects, energy efficiency, and management effects.

[0187] In some embodiments of the present disclosure, the device further includes:

[0188] A stop point acquisition module, configured to acquire stop points from the trajectory points according to the time information of the trajectory points and the spatial information of the trajectory points;

[0189] Correspondingly, the first prediction module is specifically configured to:

[0190] Predict the behavior state of the activity area based on the time information of the stop points in the movement trajectory data and the building type, and determine the behavior states of different user types;

[0191] Correspondingly, the second prediction module is specifically configured to:

[0192] Predict the electrical appliance usage behaviors of different user types in the activity area based on the behavior states of the different user types and the external environment parameters corresponding to the stop points.

[0193] In some embodiments of the present disclosure, the movement trajectory data further includes the spatial information of the trajectory points; then the stop point acquisition module includes:

[0194] A first acquisition unit, configured to, for any current trajectory point in the movement trajectory data, acquire the neighboring trajectory points of the current trajectory point according to the spatial information of the current trajectory point and the spatial information of other trajectory points;

[0195] A first determination unit, configured to, if the number of the neighboring trajectory points is greater than or equal to a first quantity threshold, use the current trajectory point as the current core point;

[0196] A second determination unit, configured to, if it is determined based on the time information of the current core point that the residence time of the current core point within a preset time period is greater than a preset time threshold, use the current core point as the current stop point;

[0197] A third determination unit, configured to update the current trajectory point to the next trajectory point, and determine the next stop point based on the time information and spatial information of the next trajectory point until all the stop points are acquired from the trajectory points.

[0198] In some embodiments of the present disclosure, the apparatus further includes:

[0199] A clustering module, configured to cluster the trajectory points in the movement trajectory data according to the distribution density of the trajectory points in the movement trajectory data to obtain at least one clustering cluster;

[0200] A noise point removal module, configured to acquire the trajectory points that do not belong to the at least one clustering cluster from the movement trajectory data, and remove the trajectory points that do not belong to the at least one clustering cluster as noise points;

[0201] Correspondingly, the first prediction module is specifically configured to:

[0202] Predict the behavior states of the activity area based on the building type and the time information of the trajectory points with noise points removed, and determine the behavior states of the different user types;

[0203] Correspondingly, the second prediction module is specifically configured to:

[0204] Predict the electrical appliance usage behaviors of different user types in the activity area according to the behavior states of the different user types and the external environment parameters corresponding to the trajectory points with noise points removed.

[0205] In some embodiments of the present disclosure, the first prediction module 620 includes:

[0206] A simulation unit, configured to simulate the time information of the trajectory point and the building type onto different nodes of a preset graph structure in the first prediction model to obtain a current graph structure;

[0207] A calculation unit, configured to calculate probability values of behavior states of different user types based on the time information and building type represented by different nodes in the current graph structure;

[0208] A second acquisition unit, configured to, for each user type, acquire the behavior state with the largest probability value from the probability values as the behavior state corresponding to each user type, and obtain the behavior states of different user types.

[0209] In some embodiments of the present disclosure, the current graph structure includes a first node abstracted from the time information, a second node abstracted from the building type, a first pointing relationship abstracted from the time information and the behavior state, and a second pointing relationship abstracted from the building type and the behavior state; then, the calculation unit is specifically configured to:

[0210] Read, from the current graph structure, the first node representing the time information, the second node representing the building type, the first pointing relationship corresponding to the time information, and the second pointing relationship corresponding to the building type;

[0211] According to the Bayesian probability calculation method, calculate probability values for the first node, the second node, the first pointing relationship, and the second pointing relationship to determine the probability values of the behavior states of different user types.

[0212] In some embodiments of the present disclosure, the apparatus further includes:

[0213] A longitude and latitude determination module, configured to map the trajectory point from a geographical space environment to a map and determine the longitude and latitude information of the trajectory point on the map;

[0214] A reading module, configured to read external environment parameters corresponding to the trajectory point from environment prediction data including a plurality of longitude and latitude information based on the longitude and latitude information of the trajectory point on the map.

[0215] In some embodiments of the present disclosure, the second prediction module 630 includes:

[0216] A model prediction unit, which uses a pre-determined behavior limit range as a constraint condition and utilizes a second prediction model to process the behavior states of different user types and the external environment parameters corresponding to the trajectory points, so as to obtain the electrical appliance usage behaviors of different user types in the activity area.

[0217] In some embodiments of the present disclosure, the model prediction unit is specifically configured to:

[0218] Based on the action decision network in the second prediction model, process the behavior states of different user types and the external environment parameters corresponding to the trajectory points to obtain multiple operation behaviors within the behavior limit range;

[0219] Based on the behavior evaluation network in the second prediction model, calculate the evaluation values corresponding to the multiple operation behaviors respectively;

[0220] According to the evaluation values corresponding to the multiple operation behaviors respectively, obtain the operation behavior with the highest evaluation value from the multiple operation behaviors as the electrical appliance usage behavior of different user types in the activity area.

[0221] In some embodiments of the present disclosure, the second prediction module 630 further includes:

[0222] A replay sample determination unit, which uses the behavior states of different user types, the external environment parameters corresponding to the trajectory points, and the electrical appliance usage behaviors of different user types as replay samples;

[0223] A replay sample storage unit, which adds the replay samples to the experience pool of the second prediction model;

[0224] A priority determination unit, which calculates the priorities of multiple replay samples to be sampled in the experience pool and determines the first priorities corresponding to the multiple replay samples to be sampled respectively;

[0225] A replay sample sampling unit, which samples multiple replay samples from the experience pool according to the first priorities corresponding to the multiple replay samples to be sampled respectively;

[0226] A loss data calculation unit, which calculates the first loss data of the action decision network and the second loss data of the behavior evaluation network based on the behavior states and electrical appliance usage behaviors included in the multiple sampled replay samples;

[0227] A model update unit, which iteratively updates the network parameters corresponding to the action decision network and the behavior evaluation network according to the first loss data and the second loss data to obtain an updated second prediction model.

[0228] In some embodiments of the present disclosure, the playback sample storage unit is specifically configured to:

[0229] Using a state value function calculation method, calculate the second priority of the playback sample;

[0230] Based on the second priority of the playback sample, add the playback sample to the experience pool of the second prediction model according to the data tree structure.

[0231] In some embodiments of the present disclosure, the simulation module 640 is specifically configured to:

[0232] For each user type, multiply the electrical appliance usage behavior of each user type by the corresponding rated power to obtain the energy consumption simulation data of each user type;

[0233] Based on the energy consumption simulation data of each user type and the proportion of each user type among the different user types, perform a weighted sum calculation to obtain the energy consumption simulation data of the activity area.

[0234] It should be noted that, Figure 6 The area power simulation device 600 based on the movement trajectory shown can execute Figures 1 to 5 each step in the method embodiments shown, and implement Figures 1 to 5 each process and effect in the method or system embodiments shown, which will not be elaborated here.

[0235] Figure 7 Fig. shows a schematic structural diagram of an area power simulation device based on the movement trajectory provided by an embodiment of the present disclosure.

[0236] As Figure 7 shown, the area power simulation device based on the movement trajectory may include a processor 701 and a memory 702 storing computer program instructions.

[0237] Specifically, the above-mentioned processor 701 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0238] The memory 702 may include a mass memory for information or instructions. By way of example and not limitation, the memory 702 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 702 may include removable or non-removable (or fixed) media. Where appropriate, the memory 702 may be internal or external to the integrated gateway device. In a particular embodiment, the memory 702 is a non-volatile solid-state memory. In a particular embodiment, the memory 702 includes a read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically alterable ROM (EAROM), or a flash memory, or a combination of two or more of these.

[0239] The processor 701 reads and executes the computer program instructions stored in the memory 702 to perform the steps of the method for area power simulation based on a moving trajectory provided by the embodiments of the present disclosure.

[0240] In one example, the area power simulation device based on a moving trajectory may further include a transceiver 703 and a bus 704. Among them, as Figure 7 shown, the processor 701, the memory 702, and the transceiver 703 are connected through the bus 704 and complete communication with each other.

[0241] The bus 704 includes hardware, software, or both. By way of example and not limitation, the bus can include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side BUS (FSB), a Hyper Transport (HT) interconnect, an Industrial Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a MicroChannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses or a combination of two or more of these. Where appropriate, the bus 704 can include one or more buses. Although the embodiments of the present application describe and illustrate specific buses, the present application contemplates any suitable bus or interconnect.

[0242] The following are embodiments of a computer-readable medium provided by the embodiments of the present disclosure. The computer-readable medium belongs to the same inventive concept as the above-described method for regional power simulation based on a moving trajectory. Details not described in detail in the embodiments of the computer-readable medium can be referred to the embodiments of the method for regional power simulation based on a moving trajectory.

[0243] This embodiment provides a medium containing computer-executable instructions that, when executed by a computer processor, are used to execute a method for regional power simulation based on a moving trajectory.

[0244] Of course, the computer-executable instructions of a medium containing computer-executable instructions provided by the embodiments of the present disclosure are not limited to the above method operations, and can also execute related operations in the method for regional power simulation based on a moving trajectory provided by any embodiment of the present disclosure. The method for regional power simulation based on a moving trajectory includes:

[0245] Obtain the movement trajectory data of different user types in the activity area, and obtain the building type of the activity area;

[0246] Based on the building type and the time information of the trajectory points in the movement trajectory data, predict the behavior state of the activity area and determine the behavior states of different user types;

[0247] According to the behavior states of different user types and the external environment parameters corresponding to the trajectory points, predict the electrical appliance usage behaviors of different user types in the activity area;

[0248] Based on the electrical appliance usage behaviors of different user types, perform energy consumption simulation on the activity area to obtain the energy consumption simulation data of the activity area.

[0249] Through the above description of the embodiments, those skilled in the art can clearly understand that the present disclosure can be implemented by means of software and necessary general-purpose hardware. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present disclosure, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. This computer software product can be stored in a computer-readable medium, such as a floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disc of a computer, etc., including several instructions for causing a computer cloud platform (which can be a personal computer, a server, or a network cloud platform, etc.) to execute the method for regional power simulation based on movement trajectory provided by each embodiment of the present disclosure.

[0250] Note that the above is only the preferred embodiment of the present disclosure and the applied technical principle. Those skilled in the art will understand that the present disclosure is not limited to the specific embodiments here. Various obvious changes, re-adjustments and substitutions can be made by those skilled in the art without departing from the protection scope of the present disclosure. Therefore, although the present disclosure has been described in detail through the above embodiments, the present disclosure is not limited to the above embodiments. Without departing from the concept of the present disclosure, it can also include more other equivalent embodiments, and the scope of the present disclosure is determined by the scope of the appended claims.

Claims

1. A regional power simulation method based on mobile trajectory, characterized in that: include: Obtain movement trajectory data of different user types in the activity area, and obtain the building type of the activity area; Based on the building type and the time information of the trajectory points in the movement trajectory data, predicting the behavior state of the activity area, and determining the behavior states of the different user types; Predicting the electrical appliance usage behaviors of different user types in the activity area according to the behavior states of the different user types and the external environment parameters corresponding to the trajectory points; Based on the electrical appliance usage behaviors of the different user types, energy consumption simulation is performed on the activity area to obtain energy consumption simulation data of the activity area; The predicting the appliance usage behaviors of different user types in the activity area according to the behavior states of the different user types and the external environment parameters corresponding to the track points includes: Taking a predetermined behavior restriction range as a constraint condition, the second prediction model is used to process the behavior states of the different user types and the external environment parameters corresponding to the trajectory points to obtain the appliance usage behaviors of the different user types in the activity area.

2. The method according to claim 1, characterized in that Before predicting the behavior state of the activity area based on the time information of the trajectory points in the movement trajectory data and the building type and determining the behavior states of the different user types, the method further includes: Acquire a stop point from the trajectory point according to the time information of the trajectory point and the spatial information of the trajectory point; Accordingly, the predicting of the behavior state of the activity area based on the time information of the trajectory points in the movement trajectory data and the building type, and determining the behavior states of the different user types include: Based on the time information of the stop points in the movement trajectory data and the building type, predicting the behavior state of the activity area, and determining the behavior states of the different user types; Accordingly, predicting the appliance usage behaviors of different user types in the activity area according to the behavior states of the different user types and the external environment parameters corresponding to the trajectory points includes: The appliance usage behaviors of the different user types in the activity area are predicted according to the behavior states of the different user types and the external environment parameters corresponding to the stay points.

3. The method according to claim 2, characterized in that The movement trajectory data also includes spatial information of the trajectory point; then, obtaining the stop point from the trajectory point according to the time information of the trajectory point and the spatial information of the trajectory point includes: For any current trajectory point in the movement trajectory data, obtaining a neighborhood trajectory point of the current trajectory point according to spatial information of the current trajectory point and spatial information of other trajectory points; If the number of the neighborhood trajectory points is greater than or equal to a first quantity threshold, taking the current trajectory point as the current core point; If it is determined based on the time information of the current core point that the stay time of the current core point within the preset time period is greater than the preset time threshold, the current core point is used as the current stay point; The current track point is updated to a next track point, and a next stop point is determined based on the time information and space information of the next track point, until all the stop points are obtained from the track points.

4. The method according to claim 1, characterized in that: Before predicting the behavior state of the activity area based on the building type and the time information of the trajectory points in the movement trajectory data to determine the behavior states of the different user types, the method further includes: Clustering the trajectory points in the movement trajectory data according to the distribution density of the trajectory points in the movement trajectory data to obtain at least one cluster; Acquire trajectory points that do not belong to the at least one cluster from the movement trajectory data, and remove the trajectory points that do not belong to the at least one cluster as noise points; Accordingly, the predicting of the behavior state of the activity area based on the building type and the time information of the trajectory point in the movement trajectory data to determine the behavior state of the different user types includes: Based on the building type and the time information of the trajectory points with noise points removed, predicting the behavior state of the activity area, and determining the behavior states of the different user types; Accordingly, predicting the appliance usage behaviors of different user types in the activity area according to the behavior states of the different user types and the external environment parameters corresponding to the trajectory points includes: The appliance usage behaviors of the different user types in the activity area are predicted according to the behavior states of the different user types and the external environment parameters corresponding to the trajectory points with noise points removed.

5. The method according to claim 1, characterized in that The predicting of the behavior state of the activity area based on the building type and the time information of the trajectory point in the movement trajectory data to determine the behavior state of the different user types includes: Simulating the time information of the trajectory point and the building type to different nodes of a preset graph structure in the first prediction model to obtain a current graph structure; Calculate the probability values ​​of the behavior states of different user types based on the time information and building types represented by different nodes in the current graph structure; For each user type, a behavior state with the largest probability value is obtained from the probability values ​​as the behavior state corresponding to each user type, thereby obtaining the behavior states of the different user types.

6. The method according to claim 5, characterized in that The current graph structure includes a first node abstracted by the time information, a second node abstracted by the building type, a first pointing relationship abstracted by the time information and the behavior state, and a second pointing relationship abstracted by the building type and the behavior state; Then, the calculation of the probability values ​​of the behavior states of different user types based on the time information and building types represented by different nodes in the current graph structure includes: Reading from the current graph structure a first node representing the time information, a second node representing the building type, a first pointing relationship corresponding to the time information, and a second pointing relationship corresponding to the building type; According to the Bayesian probability calculation method, probability values ​​of the first node, the second node, the first pointing relationship, and the second pointing relationship are calculated to determine the probability values ​​of the behavior states of the different user types.

7. The method according to claim 1, characterized in that Before predicting the appliance usage behaviors of different user types in the activity area according to the behavior states of the different user types and the external environment parameters corresponding to the trajectory points, the method further includes: Mapping the track point from the geographic space environment to a map, and determining the longitude and latitude information of the track point on the map; Based on the longitude and latitude information of the track point on the map, the external environment parameters corresponding to the track point are read from the environment prediction data containing a plurality of longitude and latitude information.

8. The method according to claim 1, characterized in that The method uses a predetermined behavior restriction range as a constraint condition, processes the behavior states of the different user types and the external environment parameters corresponding to the trajectory points using a second prediction model, and obtains the appliance usage behaviors of the different user types in the activity area, including: Based on the action decision network in the second prediction model, the behavior states of the different user types and the external environment parameters corresponding to the trajectory points are processed to obtain a plurality of operation behaviors within the behavior restriction range; Calculating the evaluation values ​​corresponding to the plurality of operation behaviors respectively based on the behavior evaluation network in the second prediction model; According to the evaluation values ​​respectively corresponding to the multiple operation behaviors, the operation behavior with the highest evaluation value is obtained from the multiple operation behaviors as the appliance usage behaviors of different user types in the activity area.

9. The method according to claim 8, characterized in that Also includes: Taking the behavior states of the different user types, the external environment parameters corresponding to the trajectory points, and the appliance usage behaviors of the different user types as playback samples, and adding the playback samples to the experience pool of the second prediction model; Performing priority calculation on a plurality of playback samples to be sampled in the experience pool, and determining first priorities respectively corresponding to the plurality of playback samples to be sampled; Sampling a plurality of playback samples from the experience pool according to the first priorities respectively corresponding to the plurality of playback samples to be sampled; Based on the behavior states and appliance usage behaviors contained in the sampled multiple playback samples, calculating the first loss data of the action decision network and the second loss data of the behavior evaluation network; According to the first loss data and the second loss data, the network parameters corresponding to the action decision network and the behavior evaluation network are iteratively updated to obtain an updated second prediction model.

10. The method according to claim 9, characterized in that The adding the playback sample to the experience pool of the second prediction model includes: Calculating the second priority of the playback sample using a state value function calculation method; Based on the second priority of the playback sample, the playback sample is added to the experience pool of the second prediction model according to the data tree structure.

11. The method according to claim 1, characterized in that: The energy consumption simulation of the activity area based on the electrical appliance usage behaviors of different user types in the activity area to obtain energy consumption simulation data of the activity area includes: For each user type, multiply the appliance usage behavior of each user type by the corresponding rated power to obtain energy consumption simulation data of each user type; The energy consumption simulation data of the activity area is obtained by performing a weighted sum calculation based on the energy consumption simulation data of each user type and the proportion of each user type in the different user types.

12. A regional power simulation device based on mobile trajectory, characterized in that: include: An acquisition module, used to acquire movement trajectory data of different user types in an activity area, and acquire the building type of the activity area; A first prediction module, configured to predict the behavior state of the activity area based on the building type and the time information of the trajectory points in the movement trajectory data, and determine the behavior states of the different user types; A second prediction module is used to predict the electrical appliance usage behavior of different user types in the activity area according to the behavior states of the different user types and the external environment parameters corresponding to the trajectory points; A simulation module, configured to simulate the energy consumption of the activity area based on the electrical appliance usage behaviors of the different user types, and obtain energy consumption simulation data of the activity area; The second prediction module includes: The model prediction unit is used to process the behavior states of the different user types and the external environment parameters corresponding to the trajectory points using a second prediction model with a predetermined behavior restriction range as a constraint condition to obtain the appliance usage behavior of the different user types in the activity area.

13. An electronic device, characterized in that: include: processor; A memory for storing executable instructions; The processor is used to read the executable instructions from the memory and execute the executable instructions to implement the method described in any one of claims 1 to 11.

14. A computer readable medium having a computer program stored thereon, characterized in that: The medium stores a computer program, and when the computer program is executed by a processor, the processor implements the method according to any one of claims 1 to 11.

Images