Intelligent campus management system and method

Abstract

The invention discloses an intelligent campus management system and method, and the method comprises the following steps: S1, collecting campus management data, and transmitting the campus management data to a cloud computing platform; s2, preprocessing the campus management data; s3, analyzing data by using an improved gating cycle unit and an exponential smoothing method; s4, performing curriculum scheduling optimization and classroom and laboratory resource allocation by adopting a super-gravitational search optimization algorithm; s5, automatically adjusting the air conditioner, the illumination, the access control authority and the monitoring strategy; s6, real-time data transmission is carried out, and the equipment response speed is optimized based on distributed calculation; s7, constructing a data sharing platform based on a block chain technology, and setting a data access permission through an intelligent contract; and S8, optimizing the campus management scheme based on reinforcement learning. The intelligent campus scheduling optimization and trend prediction system integrates the Internet of Things, artificial intelligence, a super gravitational search optimization algorithm, an improved gating loop unit and an exponential smoothing method, realizes intelligent campus scheduling optimization and trend prediction, and has the advantages of efficient optimization and intelligent management.

Description

Technical Field

[0001] The present invention relates to the field of intelligent campus management, and particularly to an intelligent campus management system and method. Background Art

[0002] With the rapid development of Internet of Things, artificial intelligence, cloud computing and blockchain technologies, intelligent campus management systems have gradually become an important means for universities, high schools and primary schools to improve management efficiency and optimize resource allocation. Traditional campus management mainly relies on manual management and decentralized information systems, such as attendance management systems, classroom reservation systems, energy consumption monitoring systems, security monitoring systems, etc. These systems usually operate independently, lacking data sharing and collaborative optimization capabilities, resulting in serious information island problems, restricting the effective utilization of data and the global optimization ability, and being unable to meet the requirements of modern campus management for intelligence, automation and high efficiency.

[0003] Most of the current intelligent campus management methods are based on simple rule settings or single optimization methods, lacking in-depth analysis and intelligent prediction capabilities for massive data. For example, existing attendance management systems usually record teachers' and students' attendance data based on RFID, fingerprint recognition or face recognition technologies, but lack intelligent analysis of attendance patterns, being unable to predict abnormal attendance situations in advance or automatically adjust attendance rules. Classroom and laboratory resource management is usually based on manual reservation or simple time conflict detection, failing to fully consider complex factors such as teachers' course scheduling requirements, students' course selection habits, and utilization rates of experimental equipment, resulting in low resource allocation efficiency. Traditional energy consumption management systems mainly rely on fixed-time rules or manual intervention, and it is difficult to dynamically adjust the operation strategies of devices such as air conditioners and lighting according to real-time data and predicted data, thus causing energy waste.

[0004] In terms of scheduling optimization, existing course scheduling and laboratory arrangement methods are mainly based on manual experience, heuristic rules or simple optimization algorithms, such as genetic algorithms, particle swarm optimization, etc. These methods have certain applicability in dealing with small-scale scheduling problems, but when facing large-scale, multi-constrained and dynamically changing campus scheduling problems, it is often difficult to find the global optimal solution, and the computational complexity is relatively high. In addition, existing scheduling methods generally ignore the consideration of future data trends, such as future classroom utilization rates, changes in energy consumption demands, teachers' and students' attendance patterns, etc., resulting in the scheduling plan possibly not conforming to the actual situation during the execution process, affecting management efficiency and resource utilization rates.

[0005] Data security and sharing issues are also a major shortcoming of existing intelligent campus management systems. Current campus management systems usually adopt a centralized database to store sensitive data such as students' academic records, exam scores, and financial transaction records, which pose risks such as data leakage, tampering, and opaque access permissions. In addition, due to the lack of an effective data sharing mechanism between different systems, teachers' and students' information is stored repeatedly on multiple platforms, resulting in poor data consistency, low query efficiency, and difficulty in achieving cross-departmental collaborative management. Some universities have tried to introduce a cloud computing-based storage architecture to improve data availability, but still cannot fundamentally solve the data security problem.

[0006] Existing campus management systems still have obvious deficiencies in intelligent optimization and adaptive adjustment. Current management solutions often rely on fixed rules and parameters and are difficult to adaptively optimize based on historical data. For example, the operation strategies of air conditioning and lighting systems usually rely on fixed-time scheduling and cannot be intelligently adjusted according to real-time environmental data, historical energy consumption trends, and current classroom usage, resulting in increased energy waste and equipment wear. In addition, the teaching management solution lacks the ability of intelligent optimization and cannot continuously optimize the scheduling strategy through reinforcement learning methods to adapt to the dynamic changes of the campus environment. Summary of the Invention

[0007] An object of the present invention is to propose an intelligent campus management system and method. The present invention integrates the Internet of Things, artificial intelligence, super-gravitational search optimization algorithm, improved gated recurrent unit, and exponential smoothing method to realize intelligent collection of campus data, trend prediction, and resource optimization scheduling. The improved gated recurrent unit is used to extract data features, combined with the exponential smoothing method to predict attendance, energy consumption, and resource usage, and the super-gravitational search optimization algorithm is used to optimize course scheduling and resource allocation in a multi-constraint environment to improve scheduling efficiency and resource utilization rate.

[0008] An intelligent campus management method according to an embodiment of the present invention includes the following steps:

[0009] S1. Collect campus management data and transmit the campus management data to the cloud computing platform through Internet of Things devices and wireless networks;

[0010] S2. Preprocess the campus management data to generate preprocessed data and store the preprocessed data in the cloud database. The key data in the preprocessed data is encrypted and stored using blockchain technology, and access permission management is implemented based on smart contracts;

[0011] S3. Analyze the preprocessed data using an improved gated recurrent unit and exponential smoothing method to predict future attendance rates, energy consumption demands, classroom and laboratory usage, and generate prediction data;

[0012] S4. Use the super-gravitational search optimization algorithm to optimize course scheduling, allocate classroom and laboratory resources, initialize the scheduling population, construct a fitness function based on course time constraints, teacher availability, classroom capacity, and experimental equipment requirements, calculate the gravitational fitness, use the prediction data as the constraints for scheduling optimization, eliminate some scheduling plans using the black hole search mechanism, and adjust the course schedule and classroom arrangements to finally obtain an optimized course scheduling plan and resource allocation plan;

[0013] S5. Execute the optimized plan and automatically adjust the air conditioning, lighting, access control permissions, and monitoring strategies based on the intelligent control system;

[0014] S6. Use 5G or WiFi6 network for real-time data transmission, optimize the device response speed based on distributed computing, and improve the data processing efficiency by combining edge computing;

[0015] S7. Build a data sharing platform based on blockchain technology. The data sharing platform stores student status information, exam scores, and financial transaction records, and sets data access permissions through smart contracts to support cross-departmental collaborative management;

[0016] S8. Optimize the campus management plan based on reinforcement learning, update the improved gated recurrent unit by combining historical campus management data, dynamically adjust the control parameters, and optimize the campus management plan.

[0017] Optionally, the campus management data includes course arrangement data, classroom usage data, laboratory resource data, teacher and student attendance data, energy consumption data, and security monitoring data. The preprocessing includes feature extraction, data format conversion, missing value filling, outlier detection, data deduplication, and standardization. The key data includes student status information, exam scores, and financial transaction records.

[0018] Optionally, the specific steps of S3 include:

[0019] S31. Extract features from the preprocessed data based on the improved gated recurrent unit, perform dilated causal convolution processing on the input sequence using a temporal convolutional network to extract long-term dependence features:

[0020]

[0021] where, represents the output of the temporal convolutional network at time step t, k represents the convolutional kernel size, W i represents the convolutional kernel weight, d represents the dilation coefficient, and x t-d·i represents the input data at time step t - d·i;

[0022] S32. Use the output of the temporal convolutional network as the input x of the gated recurrent unitt ' , and calculate the attention weights at time step t:

[0023]

[0024] e t = v B tanh(Wh t-1 + Ux t '+ b);

[0025] Among them, α t represents the attention weight at time step t, exp represents the natural exponential function, e t represents the weight score at time step t, T represents the total number of time steps, W and U represent weight matrices, h t-1 represents the hidden state of the previous time step, x t ' represents the input data of the current time step, b represents the bias term, B represents the transpose operation, v represents the attention layer parameter vector, and tanh represents the hyperbolic tangent activation function;

[0026] And calculate the hidden state after attention weighting:

[0027] h t ′ = α t h t ;

[0028] Among them, h t ′ represents the hidden state after attention weighting;

[0029] S33. Calculate the hidden state of each time step t based on the improved gated recurrent unit:

[0030] z t = σ(W z x t '+ U z h t-1 ′ + b z );

[0031] r t = σ(W r x t '+ U r h t-1 ′ + b r );

[0032]

[0033] Among them, z t represents the update gate, which controls the update ratio between the state h t-1 ′ of the previous time step and the current candidate state rt represents the reset gate, controlling the influence degree of the previous time step state on the current state calculation, W z , U z , W r , U r , W h , and U h represents the weight matrix, b z , b r and b h represents the bias term, σ represents the Sigmoid activation function, ⊙ represents element-wise multiplication, x t ' represents the input data at the current time step, h t-1 ′ represents the hidden state after attention weighting at time step t - 1, tanh represents the hyperbolic tangent activation function, represents the current candidate state;

[0034] S34. Analyze time series data using the exponential smoothing method, predict the attendance rate, energy consumption demand, classroom and laboratory usage using the triple exponential smoothing model, calculate the predicted values for future time steps, and generate prediction data:

[0035] S t = αh t ′+(1 - α)(S t-1 + b t-1 );

[0036] b t = β(S t - S t-1 )+(1 - β)b t-1 ;

[0037] T t = γ(S t - S t-m )+(1 - γ)T t-1 ;

[0038] F t+h = S t + hb t + T t ;

[0039] Among them, S t represents the smoothed value at time step t, h t ′ represents the hidden state at time step t, S t-1 represents the smoothed value at time step t - 1, b t-1 and b t represent the trend terms, m represents the cycle length, describing the calculation range of the periodic trend, T t and T t-1 represent the periodic terms, F t+hIt represents the predicted value for the next h steps, where h represents the number of steps, and α, β, and γ represent smoothing factors that control the update ratios of each item.

[0040] Optionally, S4 specifically includes:

[0041] S41. Optimize the course schedule, allocate classroom and laboratory resources based on the super-gravitational search optimization algorithm, use the predicted data as the constraint condition for schedule optimization, initialize the schedule population, and set the search space. The search space consists of all course arrangement plans, classroom and laboratory allocation plans. Define the fitness function f(X) to evaluate the optimization degree of the schedule plan, where X is the scheduling variable to be optimized, including course time, teaching teacher, classroom number, and laboratory number. The super-gravitational search optimization algorithm is an intelligent optimization algorithm based on the physical law of gravity, which combines the gravitational search algorithm and the black hole theory to find the optimal solution in a complex search space:

[0042]

[0043] Among them, f(X) represents the fitness function, O represents the total number of courses, and w1, w2, w3, w4, and w5 represent weight coefficients. represents the balance of the course time arrangement. represents the teacher availability constraint. represents the capacity fitness of the classroom and laboratory. represents the course time conflict situation, and g5(F) represents the constraint penalty term of the predicted data F on the schedule. represents the course time. represents the teaching teacher. represents the classroom number. represents the laboratory number. represents whether there is a conflict in the course time arrangement. F represents the predicted data, including attendance rate, energy consumption demand, classroom and laboratory usage.

[0044] S42. Calculate the gravitational fitness of the initial population and define the individual X i 's mass M i :

[0045]

[0046] Among them, f(X i ) represents the fitness value of the individual X i , f min represents the minimum fitness value in the current population, f max represents the maximum fitness value in the current population, and M i represents the mass of the individual X i .

[0047] S43. Calculate the gravitational force between individuals in the population based on the super-gravitational search, and define individual X i The gravitational force F j exerted on individual X ij by individual X

[0048]

[0049] is as follows: ij where F i represents the gravitational force exerted on individual X j by individual X t , G i represents the gravitational constant, and an exponential decay strategy is adopted. M i represents the mass of individual X j , M j represents the mass of individual X ij , d i represents the Euclidean distance between individual X j and individual X

[0050] in the search space, Θ represents the gravitational decay factor, and ε represents a constant to prevent the denominator from being zero;

[0051]

[0052] X i (t + 1) = X i (t) + v i (t + 1) + δ;

[0053] where v i (t + 1) represents the velocity of individual X i at iteration t + 1, v i (t) represents the velocity of individual X i at iteration t, w represents the inertia weight, φ j represents a random number used to control the degree of gravitational influence, F ij represents the gravitational force exerted on individual X i by individual X j , d ij represents the Euclidean distance between individual X i and individual X j in the search space, ε represents a constant to prevent the denominator from being zero, X i (t + 1) represents the new position of individual X i at iteration t + 1, X i (t) represents individual X iThe position at iteration t, δ represents a Gaussian random perturbation term that follows a Gaussian distribution with a mean of 0 and a variance of 1, and ζ represents the prediction data constraint penalty coefficient. It represents the gradient influence of the prediction data on the optimization objective.

[0054] S45. Use the black hole search mechanism to eliminate some mass scheduling schemes. Define the black hole fitness threshold of individual X i :

[0055] T h = f max - ξ(f max - f min );

[0056] Among them, T h represents the black hole fitness threshold, and ξ represents the black hole contraction coefficient.

[0057] When the fitness value f(X i ) of individual X i is less than the black hole fitness threshold T h , the individual is absorbed by the black hole and a new individual is regenerated:

[0058]

[0059] Among them, represents the newly generated individual, X best represents the current optimal individual, η represents the perturbation factor, and N(0,1) represents Gaussian noise to keep the newly generated individuals diverse.

[0060] S46. Adjust the course schedule and classroom arrangement. Define the perturbation change of individual X i :

[0061] ΔX i = κ·(X g - X i ) + ψ·(X rand - X i ) - ρg5(F);

[0062]

[0063] Among them, ΔX i represents the perturbation change of individual X i , X g represents the current global optimal individual, κ and ψ represent the search step adjustment parameters that combine global convergence and local exploration ability, X rand represents a randomly selected individual, represents the position of the newly calculated individual, and ρ represents the prediction data constraint weight to ensure that the scheduling optimization conforms to the prediction trend.

[0064] S47. Calculate the final fitness values of the optimized course scheduling plan and resource allocation plan. If the preset termination condition is satisfied, output the optimized course scheduling plan and resource allocation plan; otherwise, return to step S42 to continue iterative optimization, and finally generate an optimized course scheduling plan and resource allocation plan that meet the course time constraints, teacher availability, classroom capacity, and experimental equipment requirements, and store them in the cloud database.

[0065] An intelligent campus management system according to an embodiment of the present invention includes:

[0066] A data collection module, configured to collect campus management data and transmit the campus management data to the cloud computing platform through Internet of Things devices and wireless networks;

[0067] A data preprocessing module, configured to preprocess the campus management data, store the preprocessed data in the cloud database, encrypt and store the key data in the preprocessed data using blockchain technology, and implement access permission management based on a smart contract;

[0068] A data analysis and prediction module, configured to analyze the preprocessed data using an improved gated recurrent unit and exponential smoothing method, predict future attendance rates, energy consumption demands, classroom and laboratory usage, and generate prediction data;

[0069] A scheduling optimization module, configured to perform course scheduling optimization, classroom and laboratory resource allocation based on a super-gravitational search optimization algorithm, use the prediction data as a constraint condition for scheduling optimization, and adjust the course schedule and classroom arrangements;

[0070] An intelligent control module, configured to execute the optimized plan and automatically adjust air conditioners, lighting, access permissions, and monitoring strategies based on an intelligent control system;

[0071] A real-time data transmission module, configured to perform real-time data transmission using a 5G or WiFi6 network, optimize the device response speed based on distributed computing, and improve the data processing efficiency in combination with edge computing;

[0072] A data sharing module, configured to build a data sharing platform based on blockchain technology, set data access permissions through a smart contract, and support cross-departmental collaborative management;

[0073] A self-learning optimization module, configured to optimize the campus management plan based on reinforcement learning, update the improved gated recurrent unit in combination with historical campus management data, dynamically adjust control parameters, and optimize the campus management plan.

[0074] The beneficial effects of the present invention are:

[0075] First of all, through the integration and optimization of intelligent technologies, the present invention realizes the full - scale intelligence and high - efficiency of campus management. Compared with traditional campus management methods, the present invention uses Internet of Things devices to comprehensively collect campus management data and transmits it to the cloud computing platform through wireless networks to ensure the timeliness and integrity of data. The blockchain technology is used to encrypt and store key data such as student status information, examination results, and financial transactions, and access permission management is realized through smart contracts, improving data security and sharing efficiency, enabling different departments to more efficiently and transparently cooperate in managing campus resources.

[0076] Secondly, the intelligent prediction and optimization ability of the present invention enhances the reasonable scheduling of campus resources. Based on the improved gated recurrent unit and temporal convolutional network, deep feature extraction is performed on the pre - processed data, and combined with the exponential smoothing method for time - series prediction, enabling the system to accurately predict future attendance rates, energy consumption demands, classroom and laboratory usage situations. The introduction of this prediction ability not only makes the scheduling optimization forward - looking but also enables dynamic adjustment of resource allocation, reducing sudden scheduling conflicts and improving resource utilization. For example, based on the prediction of classroom usage, the course schedule can be optimized in advance to reduce the occurrence of empty classrooms or resource shortages; based on the prediction of energy consumption trends, the air - conditioning and lighting strategies can be adjusted in advance to reduce unnecessary energy waste, thereby improving energy management efficiency while ensuring the normal operation of the campus.

[0077] Thirdly, the present invention uses the super - gravitational search optimization algorithm for course scheduling optimization, classroom and laboratory resource allocation. During the scheduling optimization process, predicted data is introduced as a constraint condition, enabling the optimization process to not only consider the current resource status but also comprehensively consider future demand trends, ensuring the rationality and stability of the scheduling plan. Through the black - hole search mechanism, inferior scheduling plans are eliminated, and the course schedule and classroom arrangements are optimized, further enhancing the search ability and convergence speed of the optimization algorithm. Compared with traditional methods such as genetic algorithms and particle swarm optimization, it can find the global optimal solution more quickly in a large - scale, multi - constraint, and highly dynamic campus scheduling environment, thereby reducing scheduling conflicts and improving the overall scheduling efficiency.

[0078] In addition, the intelligent control system of the present invention realizes the automated management of the campus environment. The optimized scheduling plan can directly drive the intelligent control system to automatically adjust air - conditioning, lighting, access control permissions, and monitoring strategies, ensuring the comfort, security, and energy utilization rate of the campus environment. The real - time data transmission module uses 5G or WiFi6 networks for data transmission and combines distributed computing to optimize the device response speed. At the same time, edge computing is used to reduce the latency of remote data interaction, improving the real - time performance and stability of the system. This intelligent control method transforms campus management from a passive response mode to an active optimization mode, greatly reducing the workload of management personnel and improving campus operation efficiency.

[0079] Finally, the present invention optimizes the campus management solution through reinforcement learning, enabling the system to continuously learn and optimize its own parameters during long-term operation. By combining historical campus management data to update and improve the gated recurrent unit, it realizes dynamic adjustment of control parameters, ensuring that the management strategy can be continuously optimized as the actual situation changes. This adaptive optimization ability enables the system to maintain high-efficiency operation in the long term, adapt to the needs of different campus environments, and avoid the problem of rigid management strategies caused by fixed parameters in traditional systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, but do not constitute a limitation to the present invention. In the drawings:

[0081] Figure 1 is a flowchart of an intelligent campus management method proposed by the present invention;

[0082] Figure 2 is a schematic diagram of optimizing course scheduling and allocating classroom and laboratory resources by using a super-gravitational search optimization algorithm for an intelligent campus management method proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0083] Now, the present invention will be further described in detail with reference to the drawings. These drawings are all simplified schematic diagrams, only showing the basic structure of the present invention in a schematic manner, so they only show the components related to the present invention.

[0084] Refer to Figure 1 and Figure 2 , an intelligent campus management method, including the following steps:

[0085] S1. Collect campus management data and transmit the campus management data to the cloud computing platform through Internet of Things devices and wireless networks;

[0086] S2. Preprocess the campus management data to generate preprocessed data and store the preprocessed data in the cloud database. The key data in the preprocessed data is encrypted and stored using blockchain technology, and access permission management is implemented based on smart contracts;

[0087] S3. Analyze the preprocessed data using an improved gated recurrent unit and exponential smoothing method to predict future attendance rates, energy consumption requirements, classroom and laboratory usage conditions, and generate prediction data;

[0088] S4. Use the super-gravitational search optimization algorithm to optimize the course schedule, allocate classroom and laboratory resources, initialize the scheduling population, construct a fitness function based on course time constraints, teacher availability, classroom capacity, and experimental equipment requirements, calculate the gravitational fitness, use the prediction data as a constraint for scheduling optimization, eliminate some scheduling plans using the black hole search mechanism, and adjust the course schedule and classroom arrangements. Finally, obtain the optimized course scheduling plan and resource allocation plan;

[0089] S5. Execute the optimized plan and automatically adjust the air conditioning, lighting, access control permissions, and monitoring strategies based on the intelligent control system;

[0090] S6. Use 5G or WiFi6 network for real-time data transmission, optimize the device response speed based on distributed computing, and improve the data processing efficiency by combining edge computing;

[0091] S7. Build a data sharing platform based on blockchain technology. The data sharing platform stores student status information, exam scores, and financial transaction records, and sets data access permissions through smart contracts to support cross-departmental collaborative management;

[0092] S8. Optimize the campus management plan based on reinforcement learning, update the improved gated recurrent unit by combining historical campus management data, dynamically adjust the control parameters, and optimize the campus management plan.

[0093] In this embodiment, the campus management data includes course arrangement data, classroom usage data, laboratory resource data, teacher and student attendance data, energy consumption data, and security monitoring data. The preprocessing includes feature extraction, data format conversion, missing value filling, outlier detection, data deduplication, and standardization. The key data includes student status information, exam scores, and financial transaction records.

[0094] In this embodiment, the specific steps of S3 are as follows:

[0095] S31. Extract features from the preprocessed data based on the improved gated recurrent unit, perform dilated causal convolution processing on the input sequence using a temporal convolutional network, and extract long-term dependence features:

[0096]

[0097] Among them, represents the output of the temporal convolutional network at time step t, k represents the convolutional kernel size, W i represents the convolutional kernel weight, d represents the dilation coefficient, and x t-d·i represents the input data at time step t - d·i;

[0098] S32. Take the output As the input x of the gated recurrent unit t ', and calculate the attention weights at time step t:

[0099]

[0100] e t = v B tanh(Wh t-1 + Ux t '+ b);

[0101] Among them, α t represents the attention weight at time step t, exp represents the natural exponential function, e t represents the weight score at time step t, T represents the total number of time steps, W and U represent weight matrices, h t-1 represents the hidden state of the previous time step, x t ' represents the input data of the current time step, b represents the bias term, B represents the transpose operation, v represents the attention layer parameter vector, and tanh represents the hyperbolic tangent activation function;

[0102] And calculate the attention-weighted hidden state:

[0103] h t ′ = α t h t ;

[0104] Among them, h t ′ represents the attention-weighted hidden state;

[0105] S33. Calculate the hidden state at each time step t based on the improved gated recurrent unit:

[0106] z t = σ(W z x t '+ U z h t-1 ′ + b z );

[0107] r t = σ(W r x t '+ U r h t-1 ′ + b r );

[0108]

[0109] Among them, z t represents the update gate, which controls the update ratio between the state h t-1 ′ of the previous time step and the current candidate state r tRepresents the reset gate, controlling the influence degree of the previous time step state on the current state calculation, W z , U z , W r , U r , W h , and U h Represents the weight matrix, b z , b r and b h Represents the bias term, σ represents the Sigmoid activation function, ⊙ represents element-wise multiplication, x t ' Represents the input data of the current time step, h t-1 ′ represents the hidden state after attention weighting at time step t - 1, tanh represents the hyperbolic tangent activation function, Represents the current candidate state;

[0110] S34. Analyze time series data using the exponential smoothing method, predict the attendance rate, energy consumption demand, classroom and laboratory usage using the triple exponential smoothing model, calculate the predicted values for future time steps, and generate prediction data:

[0111] S t = αh t ′+(1 - α)(S t-1 + b t-1 );

[0112] b t = β(S t - S t-1 )+(1 - β)b t-1 ;

[0113] T t = γ(S t - S t-m )+(1 - γ)T t-1 ;

[0114] F t+h = S t + hb t + T t ;

[0115] Among them, S t represents the smoothed value at time step t, h t ′ represents the hidden state at time step t, S t-1 represents the smoothed value at time step t - 1, b t-1 and b t represent the trend terms, m represents the cycle length, describing the calculation range of the periodic trend, T t and T t-1 represent the cycle terms, F t+hIt represents the predicted value for the next h steps, where h represents the number of steps, and α, β, and γ represent smoothing factors that control the update ratios of each item.

[0116] In this embodiment, S4 specifically includes:

[0117] S41. Optimize the course schedule and allocate classroom and laboratory resources based on the super-gravitational search optimization algorithm. Use the predicted data as the constraint conditions for schedule optimization. Initialize the schedule population and set the search space, which consists of all course arrangement plans, classroom and laboratory allocation plans. Define the fitness function f(X) to evaluate the optimization degree of the schedule plan, where X is the scheduling variable to be optimized, including course time, instructor, classroom number, and laboratory number. The super-gravitational search optimization algorithm is an intelligent optimization algorithm based on the physical law of gravity, combining the gravitational search algorithm and the black hole theory, and is used to find the optimal solution in a complex search space:

[0118]

[0119] Among them, f(X) represents the fitness function, O represents the total number of courses, and w1, w2, w3, w4, and w5 represent weight coefficients. represents the balance of course time arrangements. represents the instructor availability constraint. represents the capacity fitness of the classroom and laboratory. represents the course time conflict situation, and g5(F) represents the constraint penalty term of the predicted data F on the schedule. represents the course time. represents the instructor. represents the classroom number. represents the laboratory number. represents whether there is a conflict in the course time arrangement. F represents the predicted data, including attendance rate, energy consumption demand, and classroom and laboratory usage.

[0120] S42. Calculate the gravitational fitness of the initial population and define the individual X i 's mass M i :

[0121]

[0122] Among them, f(X i ) represents the fitness value of the individual X i . f min represents the minimum fitness value in the current population, and f max represents the maximum fitness value in the current population. M i represents the mass of the individual X i .

[0123] S43. Calculate the gravitational force between individuals in the population based on the super-gravitational search, and define individual X i The gravitational force F j exerted on individual X ij :

[0124]

[0125] Among them, F ij represents the gravitational force exerted on individual X i by individual X j , G t represents the gravitational constant, adopting an exponential decay strategy, M i represents the mass of individual X i , M j represents the mass of individual X j , d ij represents the Euclidean distance between individual X i and individual X j in the search space, Θ represents the gravitational decay factor, and ε represents a constant to prevent the denominator from being zero;

[0126] S44. Update the position of the individual based on the calculated gravitational value, and define the velocity update rule:

[0127]

[0128] X i (t + 1) = X i (t) + v i (t + 1) + δ;

[0129] Among them, v i (t + 1) represents the velocity of individual X i at iteration t + 1, v i (t) represents the velocity of individual X i at iteration t, w represents the inertia weight, φ j represents a random number used to control the degree of gravitational influence, F ij represents the gravitational force exerted on individual X i by individual X j , d ij represents the Euclidean distance between individual X i and individual X j in the search space, ε represents a constant to prevent the denominator from being zero, X i (t + 1) represents the new position of individual X i at iteration t + 1, X i (t) represents individual X iThe position at iteration time t, δ represents a Gaussian random perturbation term, which follows a Gaussian distribution with a mean of 0 and a variance of 1, and ζ represents the penalty coefficient for predicting data constraints. It represents the gradient influence of the predicted data on the optimization objective.

[0130] S45. Use the black hole search mechanism to eliminate some quality scheduling schemes, and define the black hole fitness threshold of individual X i :

[0131] T h = f max - ξ(f max - f min );

[0132] Among them, T h represents the black hole fitness threshold, and ξ represents the black hole contraction coefficient.

[0133] When the fitness value f(X i ) of individual X i is less than the black hole fitness threshold T h , the individual is absorbed by the black hole and a new individual is regenerated:

[0134]

[0135] Among them, represents the newly generated individual, X best represents the current optimal individual, η represents the perturbation factor, and N(0,1) represents Gaussian noise, so that the newly generated individuals maintain diversity.

[0136] S46. Adjust the course schedule and classroom arrangement, and define the perturbation change amount of individual X i :

[0137] ΔX i = κ·(X g - X i ) + ψ·(X rand - X i ) - ρg5(F);

[0138]

[0139] Among them, ΔX i represents the perturbation change amount of individual X i , X g represents the current global optimal individual, κ and ψ represent the search step adjustment parameters, combining global convergence and local exploration capabilities, X rand represents a randomly selected individual, represents the calculated position of the new individual, and ρ represents the predicted data constraint weight to ensure that the scheduling optimization conforms to the predicted trend.

[0140] S47. Calculate the final fitness values of the optimized course scheduling plan and resource allocation plan. If the preset termination condition is satisfied, output the optimized course scheduling plan and resource allocation plan; otherwise, return to step S42 to continue iterative optimization until an optimized course scheduling plan and resource allocation plan that meet the course time constraints, teacher availability, classroom capacity, and experimental equipment requirements are generated and stored in the cloud database.

[0141] An intelligent campus management system, comprising:

[0142] A data collection module, configured to collect campus management data and transmit the campus management data to a cloud computing platform through Internet of Things devices and a wireless network;

[0143] A data preprocessing module, configured to preprocess the campus management data, store the preprocessed data in a cloud database, encrypt and store the key data in the preprocessed data using blockchain technology, and implement access right management based on a smart contract;

[0144] A data analysis and prediction module, configured to analyze the preprocessed data using an improved gated recurrent unit and exponential smoothing method, predict future attendance rates, energy consumption demands, classroom and laboratory usage conditions, and generate prediction data;

[0145] A scheduling optimization module, configured to perform course scheduling optimization, classroom and laboratory resource allocation based on a super-gravitational search optimization algorithm, use the prediction data as constraints for scheduling optimization, and adjust the course schedule and classroom arrangements;

[0146] An intelligent control module, configured to execute the optimized plan and automatically adjust air conditioners, lighting, access control permissions, and monitoring strategies based on an intelligent control system;

[0147] A real-time data transmission module, configured to perform real-time data transmission using a 5G or WiFi6 network, optimize the device response speed based on distributed computing, and improve data processing efficiency by combining edge computing;

[0148] A data sharing module, configured to build a data sharing platform based on blockchain technology, set data access permissions through a smart contract, and support cross-departmental collaborative management;

[0149] A self-learning optimization module, configured to optimize the campus management plan based on reinforcement learning, update the improved gated recurrent unit by combining historical campus management data, dynamically adjust control parameters, and optimize the campus management plan.

[0150] Embodiment 1:

[0151] To verify the feasibility of the present invention in implementation, the present invention is applied to the intelligent campus management system of a certain university, and experiments are carried out in aspects such as course scheduling, classroom and laboratory resource allocation, and energy consumption optimization, to evaluate the effects of the present invention in improving resource utilization rate, reducing energy consumption, and optimizing scheduling efficiency. The university covers an area of about 1,200 mu, with a total of 35,000 undergraduate and graduate students, 20 teaching buildings, 150 laboratories, and 450 classrooms. More than 8,000 courses are offered on average each semester. The existing campus management system adopts the traditional rule-based scheduling method, which fails to combine real-time data with predictive analysis, resulting in a low utilization rate of classroom resources, frequent course conflicts, and a large room for optimization in energy consumption management.

[0152] In the application in this university, campus management data is first collected through Internet of Things devices, including the attendance records of teachers and students, course arrangements, the usage of classrooms and laboratories, energy consumption data, and security monitoring information. These data are transmitted to the cloud computing platform through wireless networks and preprocessed, including data format conversion, outlier detection, data deduplication, feature extraction, and standardization processing. At the same time, key data such as student status information, exam scores, and financial data are encrypted and stored using blockchain technology to ensure the security and traceability of the data.

[0153] In the data analysis stage, an improved gated recurrent unit and temporal convolutional network are used to extract features from the data, and the exponential smoothing method is combined to predict future attendance rates, classroom and laboratory usage, and energy consumption requirements. For example, for the attendance data within a week, it is analyzed that the attendance rates of certain courses are low at specific times, the usage rates of laboratory equipment far exceed the load during some peak periods, while the usage frequencies of some classrooms are low. The exponential smoothing method prediction shows that in winter, the air-conditioning energy consumption of some large-classrooms can be 15% higher than the actual demand, and some laboratories still maintain high energy consumption at night.

[0154] Based on the above prediction data, the hyper-gravitational search optimization algorithm is used to optimize course scheduling and resource allocation. First, the search space is set, covering all feasible course arrangement plans, teacher availability, classroom and laboratory capacities, and equipment requirements and other variables. Subsequently, the fitness value is calculated through the hyper-gravitational search optimization algorithm, low-quality scheduling plans are eliminated based on the black hole search mechanism, and further optimization is carried out through the space-time bending search strategy.

[0155] Table 1 Comparison table of experimental data

[0156] Index Before optimization After optimization Improvement rate (%) Total number of courses 8250 8250 - Number of course conflicts 560 110 -80.4％ Course conflict rate (%) 6.79 1.33 -80.4％ Classroom utilization rate (%) 72.4 86.1 +18.9％ Laboratory utilization rate (%) 68.9 79.5 +15.4％ Total power consumption (MWh) 4.72 3.87 -18.0％ Classroom air-conditioning energy consumption (MWh) 2.15 1.72 -20.0％ Laboratory equipment energy consumption (MWh) 1.26 1.01 -19.8％ Nighttime vacant energy consumption (MWh) 0.89 0.67 -25.0％ Course scheduling time (hours) 26 9 -65.4％ Equipment response time (seconds) 3.5 1.2 -65.7％ Data processing speed (records / second) 1200 2100 +75.0％

[0157] The comparison table of experimental data clearly shows the optimization effect of the present invention in intelligent campus management. From the perspective of course scheduling, after optimization, the number of course conflicts has decreased from 560 times to 110 times, and the conflict rate has dropped from 6.79% to 1.33%, a reduction of 80.4%, improving the rationality of course arrangements and teacher availability. At the same time, the utilization rate of classrooms has increased from 72.4% to 86.1%, and the utilization rate of laboratories has risen from 68.9% to 79.5%, an increase of 18.9% and 15.4% respectively, indicating that the optimized scheduling scheme distributes teaching resources more evenly, effectively reducing space waste and improving the usage efficiency of classrooms and laboratories.

[0158] In terms of energy consumption management, the total electricity consumption of the campus after optimization has decreased from 4.72 MWh to 3.87 MWh, with an overall energy saving of 18%. Among them, the energy consumption of classroom air conditioners has decreased from 2.15 MWh to 1.72 MWh, a reduction of 20%; the energy consumption of laboratory equipment has dropped from 1.26 MWh to 1.01 MWh, a decrease of 19.8%; and the energy consumption during nighttime vacancy has decreased from 0.89 MWh to 0.67 MWh, a reduction of 25%. This shows that the intelligent control system can dynamically adjust the operating status of equipment such as air conditioners and lighting by combining prediction data and optimized scheduling, effectively reducing energy waste while ensuring the comfort of the teaching environment.

[0159] In terms of management efficiency, the course scheduling time after optimization has been shortened from 26 hours to 9 hours, a reduction of 65.4%, indicating that the scheduling scheme based on the super-gravitational search optimization algorithm has improved the automation and computational efficiency of scheduling. At the same time, the device response time has been shortened from 3.5 seconds to 1.2 seconds, a reduction of 65.7%, indicating that the system has optimized data transmission through 5G or WiFi6 networks and improved the real-time response ability of devices by combining edge computing. In addition, the data processing speed has increased from 1,200 records per second to 2,100 records per second, an increase of 75%, demonstrating the advantage of the combination of the cloud computing platform and the optimization algorithm in high-concurrency data processing, enabling the system to process massive campus management data more efficiently.

[0160] Generally speaking, the present invention has achieved optimization in course scheduling, resource allocation, energy consumption management, and system operation efficiency. Through real-time data collection by the Internet of Things, predictive analysis using the improved gated recurrent unit and exponential smoothing method, scheduling optimization using the super-gravitational search optimization algorithm, and adaptive adjustment of the intelligent control system, the overall intelligent level of campus management has been improved. The optimized system can arrange teaching resources more reasonably, reduce energy consumption, and improve management efficiency, providing a more efficient, energy-saving, and intelligent management mode for universities.

[0161] The above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent substitutions or changes, shall be covered by the protection scope of the present invention.

Claims

1. An intelligent campus management method, characterized in that, It includes the following steps: S1. Collect campus management data and transmit the campus management data to the cloud computing platform through Internet of Things devices and wireless networks; S2. Preprocess the campus management data to generate preprocessed data, and store the preprocessed data in the cloud database. The key data in the preprocessed data is encrypted and stored using blockchain technology, and access permission management is implemented based on smart contracts; S3. Analyze the preprocessed data using an improved gated recurrent unit and exponential smoothing method to predict future attendance rates, energy consumption demands, classroom and laboratory usage, and generate prediction data; S4. Use the super-gravitational search optimization algorithm for course scheduling optimization and classroom and laboratory resource allocation. Initialize the scheduling population, and construct a fitness function based on course time constraints, teacher availability, classroom capacity, and experimental equipment requirements. Calculate the gravitational fitness, use the prediction data as a constraint condition for scheduling optimization, and use the black hole search mechanism to eliminate some scheduling plans, and adjust the course schedule and classroom arrangements. Finally, obtain the optimized course scheduling plan and resource allocation plan; S5. Execute the optimized plan, and automatically adjust the air conditioning, lighting, access control permissions, and monitoring strategies based on the intelligent control system; S6. Use 5G or WiFi6 networks for real-time data transmission, optimize the device response speed based on distributed computing, and combine edge computing to improve data processing efficiency; S7. Build a data sharing platform based on blockchain technology. The data sharing platform stores student status information, exam scores, and financial transaction records, and sets data access permissions through smart contracts to support cross-departmental collaborative management; S8. Optimize the campus management plan based on reinforcement learning, update the improved gated recurrent unit in combination with historical campus management data, dynamically adjust control parameters, and optimize the campus management plan.

2. The intelligent campus management method according to claim 1, wherein The campus management data includes course arrangement data, classroom usage data, laboratory resource data, teacher and student attendance data, energy consumption data, and security monitoring data. The preprocessing includes feature extraction, data format conversion, missing value filling, outlier detection, data deduplication, and standardization. The key data includes student status information, exam scores, and financial transaction records.

3. The intelligent campus management method according to claim 1, wherein The specific content of S3 includes: S31. Extract features from the preprocessed data based on the improved gated recurrent unit, and perform dilated causal convolution processing on the input sequence using a temporal convolutional network to extract long-term dependence features: Among them, represents the output of the temporal convolutional network at time step t, k represents the convolutional kernel size, and W i represents the convolutional kernel weight, d represents the dilation coefficient, and x t-d·i represents the input data at time step t - d·i; S32. Use the output of the temporal convolutional network as the input \(x\) of the gated recurrent unit t ', and calculate the attention weights at time step \(t\): e t = v B tanh(Wh t-1 + Ux t '+ b); where, α t represents the attention weight at time step t, exp represents the natural exponential function, and e t represents the weight score at time step t, T represents the total number of time steps, W and U represent weight matrices, and h t-1 represents the hidden state of the previous time step, x t ' represents the input data at the current time step, b represents the bias term, B represents the transpose operation, v represents the attention layer parameter vector, and tanh represents the hyperbolic tangent activation function; And calculate the attention-weighted hidden state: h t ' = α t h t ; Among them, h t ' represents the hidden state after attention weighting; S33. Calculate the hidden state at each time step t based on the improved gated recurrent unit: z t = σ(W z x t '+ U z h t-1 '+ b z ); r t = σ(W r x t '+ U r h t-1 '+ b r ); Among them, z t represents the update gate that controls the state h t-1 ' at the previous time step and the update ratio between the current candidate state r t represents the reset gate that controls the influence degree of the state at the previous time step on the calculation of the current state, W z 、U z 、W r 、U r 、W h 、and U h represent weight matrices, b z 、b r and b h represent bias terms, σ represents the Sigmoid activation function, ⊙ represents element-wise multiplication, x t ' represents the input data at the current time step, h t-1 ' represents the hidden state weighted by attention at time step t - 1, tanh represents the hyperbolic tangent activation function, represents the current candidate state; S34. Analyze the time series data using the exponential smoothing method, use the triple exponential smoothing model to predict the attendance rate, energy consumption demand, classroom and laboratory usage, calculate the predicted values at future time steps, and generate prediction data: S t = αh t '+(1 - α)(S t-1 + b t-1 ); b t = β(S t - S t-1 ) + (1 - β)b t-1 ; T t = γ(S t - S t-m ) + (1 - γ)T t-1 ; F t+h = S t + hb t + T t ; Among them, S t represents the smoothed value at time step t, h t ' represents the hidden state at time step t, S t-1 represents the smoothed value at time step t-1, b t-1 and b t represent the trend terms, m represents the cycle length, describing the calculation range of the periodic trend, T t and T t-1 represent the periodic terms, F t+h represents the predicted value for the next h steps, h represents the number of steps, and α, β, and γ represent the smoothing factors, controlling the update ratio of each term.

4. The intelligent campus management method according to claim 1, wherein The specific content of S4 includes: S41. Optimize the course schedule and allocate classroom and laboratory resources based on the super-gravitational search optimization algorithm. Use the predicted data as the constraints for schedule optimization. Initialize the schedule population and set the search space, which consists of all course arrangement plans, classroom and laboratory allocation plans. Define the fitness function f(X) to evaluate the optimization degree of the schedule plan, where X is the scheduling variable to be optimized, including course time, instructor, classroom number, and laboratory number. The super-gravitational search optimization algorithm is an intelligent optimization algorithm based on the physical law of gravity, combining the gravitational search algorithm and the black hole theory, and is used to find the optimal solution in a complex search space: Among them, f(X) represents the fitness function, O represents the total number of courses, w1, w2, w3, w4, and w5 represent the weight coefficients, represents the balance of the course schedule, represents the teacher availability constraint, represents the capacity fitness of the classroom and the laboratory, represents the course time conflict situation, and g5(F) represents the constraint penalty term of the prediction data F on the scheduling, represents the course time, represents the teaching teacher, represents the classroom number, represents the laboratory number, represents whether there is a conflict in the course schedule, and F represents the prediction data, including the attendance rate, energy consumption demand, classroom and laboratory usage; S42. Calculate the gravitational fitness of the initial population and define the mass M of individual X i of i : Among them, f(X i ) represents the fitness value of individual X i , f min represents the minimum fitness value in the current population, f max represents the maximum fitness value in the current population, M i represents the mass of individual X i . S43. Calculate the gravitational force between individuals in the population based on super-gravitational search, and define individual X i subject to individual X j The gravitational force F ij acted on is: Among them, F ij represents the gravitational force exerted on individual X i by individual X j ; G t represents the gravitational constant, and an exponential decay strategy is adopted. M i represents the mass of individual X i ; M j represents the mass of individual X j ; d ij represents the Euclidean distance between individual X i and individual X j in the search space. Θ represents the gravitational decay factor, and ε represents a constant to prevent the denominator from being zero; S44. Update the position of the individual based on the calculated gravitational value, and define the velocity update rule: X i (t + 1) = X i (t) + v i (t + 1) + δ; Among them, v i (t + 1) represents the velocity of individual X i at the (t + 1)-th iteration, v i (t) represents the velocity of individual X i at the t-th iteration, w represents the inertia weight, φ j represents a random number used to control the degree of gravitational influence, F ij represents the gravitational force exerted on individual X i by individual X j The gravitational force, d ij represents individual X i and individual X j in the Euclidean distance in the search space, ε represents a constant to prevent the denominator from being zero, X i (t + 1) represents the new position of individual X i at the (t + 1)-th iteration, X i (t) represents the position of individual X i at the t-th iteration, δ represents a Gaussian random perturbation term that follows a Gaussian distribution with a mean of 0 and a variance of 1, ζ represents the prediction data constraint penalty coefficient, represents the gradient influence of the prediction data on the optimization objective; S45. Eliminate some of the mass scheduling schemes using the black hole search mechanism, and define the individual X i The black hole fitness threshold of T h = f max - ξ(f max - f min ); Among them, T h represents the black hole fitness threshold, and ξ represents the black hole contraction coefficient; When individual X i 's fitness value f(X i ) is less than the black hole fitness threshold T h , the individual is absorbed by the black hole and a new individual is regenerated: Among them, represents the newly generated individual, X best represents the current optimal individual, η represents the perturbation factor, and N(0,1) represents Gaussian noise, which enables the newly generated individuals to maintain diversity; S46. Adjust the course schedule and classroom arrangement, and define the perturbation change of individual X i : ΔX i = κ·(X g - X i ) + ψ·(X rand - X i ) - ρg5(F); Among them, ΔX i represents the perturbation change of individual X i , X g represents the current global optimal individual, κ and ψ represent the search step size adjustment parameters, combining global convergence and local exploration capabilities, X rand represents a randomly selected individual, represents the calculated new individual position, ρ represents the prediction data constraint weight, ensuring that the scheduling optimization conforms to the prediction trend; S47. Calculate the final fitness value of the optimized course schedule plan and resource allocation plan. If the preset termination condition is met, output the optimized course schedule plan and resource allocation plan; otherwise, return to step S42 to continue iterative optimization. Finally, generate an optimized course schedule plan and resource allocation plan that meet the course time constraints, teacher availability, classroom capacity, and experimental equipment requirements, and store them in the cloud database.

5. An intelligent campus management system, which executes the intelligent campus management method according to any one of claims 1 to 4, characterized in that Including: A data collection module, which is used to collect campus management data and transmit the campus management data to the cloud computing platform through Internet of Things devices and wireless networks; A data preprocessing module, which is used to preprocess the campus management data, store the preprocessed data in the cloud database, encrypt the key data in the preprocessed data using blockchain technology, and implement access permission management based on smart contracts; A data analysis and prediction module, which is used to analyze the preprocessed data using an improved gated recurrent unit and exponential smoothing method, predict the future attendance rate, energy consumption demand, classroom and laboratory usage, and generate predicted data; A scheduling optimization module, which is used to optimize the course schedule and allocate classroom and laboratory resources based on the super-gravitational search optimization algorithm, and use the predicted data as the constraints for schedule optimization to adjust the course schedule and classroom arrangement; An intelligent control module, which is used to execute the optimized plan and automatically adjust the air conditioning, lighting, access control permissions, and monitoring strategies based on the intelligent control system; A real-time data transmission module, which is used to perform real-time data transmission using 5G or WiFi6 networks, optimize the device response speed based on distributed computing, and improve the data processing efficiency by combining edge computing; A data sharing module, which is used to build a data sharing platform based on blockchain technology, set data access permissions through smart contracts, and support cross-departmental collaborative management; A self-learning optimization module, which is used to optimize the campus management plan based on reinforcement learning, update the improved gated recurrent unit by combining historical campus management data, dynamically adjust the control parameters, and optimize the campus management plan.

Images