Intelligent campus resource dynamic optimization allocation method based on big data

Abstract

The invention relates to the field of smart campus management, and discloses a smart campus resource dynamic optimization allocation method based on big data, comprising the following steps: acquiring resource related data and demand related data of a smart campus, the resource related data including campus hardware resource data and software resource data, the demand related data comprises teacher and student use demand data and campus management demand data; and constructing a big data resource allocation analysis model, wherein the analysis model comprises a resource feature extraction module, a demand matching module and an allocation strategy generation module. By collecting related data of the smart campus, determining resource feature dimensions and demand feature dimensions by means of a big data resource allocation analysis model, and calculating a matching score through a multi-dimensional matching rule, core information of resources and demands can be associated in an omnibearing manner, accurate adaptation of the resources and the demands is realized, and the matching efficiency of the resources and the demands is improved. The problem of one-sided resource and demand matching in a traditional mode is solved, and the accuracy of resource allocation is remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of smart campus management, in particular to a smart campus resource dynamic optimization and deployment method based on big data. BACKGROUND

[0002] In the process of deepening the construction of smart campus, the efficient deployment of campus resources (covering hardware and software resources such as classrooms, laboratory instruments, academic databases, and network bandwidth) is the key support to ensure the orderly development of teaching, the smooth progress of scientific research, and the improvement of the experience of teachers and students. In the daily operation of the campus, resource demand presents multi-dimensional characteristics: teachers and students have real-time needs such as course reservation, experiment, and literature review, and the management level has planning needs such as resource maintenance, energy consumption control, and emergency dispatch, and both types of needs often change dynamically with the progress of teaching, the advancement of scientific research projects, and the arrangement of campus activities.

[0003] However, the traditional campus resource deployment mode relies on pre-set fixed rules and manual adjustment of management personnel to complete resource allocation. This mode has significant limitations and does not fully integrate the correlation information of resource characteristics and demand dimensions - for example, when allocating laboratories, only whether the instruments are idle is considered, without considering the instrument maintenance cost, the adaptability of the use scene (such as the difference between undergraduate basic experiments and graduate precise experiments), and the priority of the demand subject (such as the priority differentiation between emergency drill needs and regular teaching needs), resulting in low matching degree of resource allocation and actual demand. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a smart campus resource dynamic optimization and deployment method based on big data, which solves the problems mentioned in the background.

[0005] To achieve the above purpose, the present application realizes the following technical scheme: a smart campus resource dynamic optimization and deployment method based on big data, comprising the following steps: Obtain resource-related data and demand-related data of the smart campus, the resource-related data including campus hardware resource data and software resource data, and the demand-related data including teacher and student use demand data and campus management demand data; Construct a big data resource deployment analysis model, the analysis model including a resource feature extraction module, a demand matching module, and a deployment strategy generation module; Input the resource-related data and demand-related data into the big data resource deployment analysis model to obtain a campus resource supply and demand matching result; Determine the resource deployment priority according to the supply and demand matching result, and generate an initial resource deployment scheme based on the deployment priority; Real-time collection of execution data in the resource allocation process, dynamic adjustment of the initial resource allocation scheme according to the execution data, and obtaining an optimized campus resource allocation scheme.

[0006] Preferably, the hardware resource data includes classroom equipment data, laboratory equipment data, library collection data, and campus network equipment data. The software resource data includes teaching platform account data, academic database permission data, campus management system function data, and online course resource data. The teacher-student usage demand data includes course reservation demand data, experimental equipment usage demand data, literature borrowing demand data, and network bandwidth demand data. The campus management demand data includes resource maintenance plan data, energy consumption control demand data, space utilization rate improvement demand data, and emergency resource scheduling demand data.

[0007] Preferably, the hardware resource data is obtained by collecting classroom equipment operation state data and laboratory instrument usage data through campus Internet of Things sensors, exporting library collection borrowing and inventory data through a library management system, and extracting campus network equipment load data through a network management platform. The software resource data is obtained by collecting account activity data and course resource access data from a teaching platform backend, obtaining user permission allocation data from a database management system, and extracting function usage frequency data from a campus management system log. The teacher-student usage demand data is obtained by collecting course and equipment reservation application data submitted by teachers and students through a campus mobile terminal, recording literature borrowing request data through a library borrowing system, and collecting bandwidth usage demand data through a network traffic monitoring tool. The campus management demand data is obtained by extracting resource maintenance plan data from a campus logistics department archive, obtaining energy consumption control index data from an energy consumption monitoring system, counting regional utilization rate data from a space management platform, and exporting emergency resource scheduling demand data from an emergency management system.

[0008] Preferably, the big data resource allocation analysis model is constructed by: Determining the campus resource feature dimension and the demand feature dimension, the resource feature dimension including resource type, available time period, capacity specification, and maintenance cost, and the demand feature dimension including demand subject, demand time period, demand size, and emergency degree. Building a feature extraction module, converting resource-related data and demand-related data into standardized feature vectors through data tokenization, feature encoding, and outlier filtering processing. The demand matching module is built, and a matching rule is set, the matching rule including a coincidence degree rule of a resource available period and a demand period, an adaptation degree rule of a resource capacity and a demand scale, and a corresponding rule of a resource type and a demand type; The deployment strategy generation module is built, historical data of campus resource deployment and optimal case data are imported, and a strategy generation model is trained through a machine learning algorithm.

[0009] Preferably, the resource type includes teaching resources, scientific research resources, management resources, and life resources. The available period includes a fixed period, an elastic period, and an emergency calling period. The demand subject includes a student group, a teacher group, an administrative staff group, and a campus visitor group. The emergency degree is divided into first-level emergency, second-level emergency, and regular demand, wherein the first-level emergency corresponds to a sudden emergency scene demand, the second-level emergency corresponds to a short-term important activity demand, and the regular demand corresponds to daily teaching and office demand.

[0010] Preferably, the setting process of the matching rule includes: Successful cases and failed cases of past campus resource deployment are collected, and resource feature and demand feature parameters in the cases are extracted; The correlation coefficient of the resource feature and the demand feature in the successful cases is calculated, the correlation coefficient being obtained through Pearson correlation analysis and cosine similarity calculation; The weight of each matching rule is determined according to the correlation coefficient, the higher the correlation coefficient, the greater the weight of the corresponding matching rule, forming a final multi-dimensional matching rule system.

[0011] Preferably, inputting the resource-related data and the demand-related data into a big data resource deployment analysis model includes: The resource-related data and the demand-related data are preprocessed, the preprocessing including data deduplication, missing value filling, and data format conversion, to obtain preprocessed data; The preprocessed data is input into a feature extraction module to generate a resource feature vector and a demand feature vector; The resource feature vector and the demand feature vector are input into the demand matching module, and a matching score is calculated according to a preset matching rule, the matching score being obtained by weighted summation of each rule weight and corresponding feature matching degree; The matching score is sorted from high to low, and a campus resource supply and demand matching result list is generated.

[0012] Preferably, determining a resource deployment priority according to the supply and demand matching result includes: An evaluation index of priority is set, the evaluation index including a demand emergency degree, a resource utilization rate, a deployment cost, and a teacher-student satisfaction degree; Quantitative scoring standards are set for each evaluation index, wherein the demand urgency degree corresponds to different scores according to the emergency level, the resource utilization rate is calculated according to the ratio of the actual use time length to the total available time length, the allocation cost is counted according to the resource transportation, debugging and maintenance expenses, and the teacher and student satisfaction degree is obtained through questionnaire investigation and use feedback scoring; The comprehensive priority score of each supply-demand matching item is calculated, and the comprehensive priority score is obtained by weighting the quantitative scores of the indexes and the preset index weights; The order of resource allocation is determined according to the comprehensive priority score from high to low.

[0013] Preferably, the dynamic adjustment of the initial resource allocation scheme according to the execution data comprises: An execution data collection cycle is set, and the collection cycle comprises real-time collection, hourly collection and daily collection, wherein the real-time collection is used in the emergency resource allocation scene, and the daily collection is used in the conventional teaching resource allocation scene; Key indicators in the execution data are extracted, and the key indicators comprise the resource actual use time length, the demand satisfaction rate, the allocation delay time length and the resource failure occurrence rate; The key indicators are compared with the preset threshold value, if the key indicators exceed the preset threshold value range, the deviation reason is analyzed, and the deviation reason comprises resource allocation excess, demand prediction deviation and equipment sudden failure; According to the deviation reason, the parameters of the big data resource allocation analysis model are adjusted, the resource allocation scheme is regenerated, and dynamic optimization is realized.

[0014] Preferably, the setting process of the preset threshold value comprises: The campus resource allocation execution data in the past three years are collected, and the historical mean value and fluctuation range of each key indicator are counted; The historical mean value is corrected in combination with the campus management target and the teacher and student demand standard, wherein the resource actual use time length threshold value is set according to the course time length and the experiment time length standard, the demand satisfaction rate threshold value is not less than 90%, the allocation delay time length threshold value is not more than 30 minutes, and the resource failure occurrence rate threshold value is not higher than 5%; The corrected threshold value is verified through an expert review meeting, and the preset threshold value range of each key indicator is finally determined by adjusting the campus actual operation scene.

[0015] The present application provides a kind of based on big data's wisdom campus resource dynamic optimization allocation method.There is following beneficial effect: 1. The application can fully associate the core information of resources and demand by collecting campus hardware resource data, software resource data, teacher and student use demand data and campus management demand data, and through the multi-dimensional matching rule calculation matching score, and can realize the accurate matching of resources and demand, solve the problem of one-sidedness of resource and demand matching in the traditional mode, and significantly improve the accuracy of resource allocation.

[0016] 2. The big data resource allocation analysis model constructed by the application can collect resource allocation execution data in real time, analyze the allocation deviation combined with the preset threshold, timely adjust the model parameters and the allocation scheme, form a rapid response mechanism to the change of resource use, solve the problem of insufficient dynamic response in the traditional mode, ensure that the resource allocation scheme always fits the real-time demand of the campus, and avoid the disconnection between resource supply and demand caused by lagging scheme. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 The method steps of the application are shown in the figure. DETAILED DESCRIPTION

[0018] The technical solutions of the application will be described in detail below with reference to the drawings of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0019] Embodiment: Please refer to the attached Figure 1 The embodiment of the application provides a big data-based intelligent campus resource dynamic optimization and allocation method, which comprises the following steps: Obtain resource-related data and demand-related data of the intelligent campus, wherein the resource-related data comprises campus hardware resource data and software resource data, and the demand-related data comprises teacher and student use demand data and campus management demand data; Construct a big data resource allocation analysis model, wherein the analysis model comprises a resource feature extraction module, a demand matching module and an allocation strategy generation module; Input the resource-related data and demand-related data into the big data resource allocation analysis model to obtain a campus resource supply and demand matching result; Determine the resource allocation priority according to the supply and demand matching result, and generate an initial resource allocation scheme based on the allocation priority; Collect execution data in the resource allocation process in real time, dynamically adjust the initial resource allocation scheme according to the execution data, and obtain an optimized campus resource allocation scheme.

[0020] The hardware resource data includes classroom equipment data, laboratory equipment data, library collection data and campus network equipment data; The software resource data includes teaching platform account data, academic database permission data, campus management system function data and online course resource data; The teacher-student use demand data includes course reservation demand data, experimental equipment use demand data, literature borrowing demand data and network bandwidth demand data; The campus management demand data includes resource maintenance plan data, energy consumption control demand data, space utilization rate improvement demand data and emergency resource scheduling demand data.

[0021] Specifically, in the hardware resource data, the classroom equipment data covers the running state (such as the number of faults, the use time) and the position information of the projectors, multimedia equipment and the like, the laboratory equipment data includes key parameters such as instrument model, accuracy level, available period, the library collection data records information such as book ISBN, inventory quantity, borrowing frequency, and the campus network equipment data includes load rate, bandwidth occupancy and the like of routers and switches; In the software resource data, the teaching platform account data records the user to which the account belongs, the permission level and the active time length, the academic database permission data clearly defines the database range and download permission accessible by different user groups, the campus management system function data statistics the use frequency and operation log of each management module (such as logistics, education), and the online course resource data includes course name, play quantity, teacher-student evaluation and the like; In the teacher-student use demand data, the course reservation demand data records the course name, the required classroom specification and the reservation period, the experimental equipment use demand data includes experimental project, required instrument type and use time length, the literature borrowing demand data records the required literature name and borrowing period, and the network bandwidth demand data statistics the bandwidth use peak value of different regions and different time periods; In the campus management demand data, the resource maintenance plan data clearly defines the equipment maintenance time and maintenance type (such as daily maintenance, fault repair), the energy consumption control demand data sets the upper limit of energy consumption of each region (such as classroom power consumption, air conditioner use standard), the space utilization rate improvement demand data includes the utilization rate target value of each region (such as laboratory, conference room), and the emergency resource scheduling demand data records the required resource type and scheduling time limit of sudden scenarios (such as fire, equipment failure); Through the detailed classification of the resource and demand data, the accuracy of data collection is realized, and the analysis deviation caused by fuzzy data classification is avoided; each data category comprehensively covers the core scenarios of campus operation, provides a data basis with fine granularity for subsequent model construction and matching analysis, and ensures that the deployment scheme can fit the resource and demand characteristics of different scenarios.

[0022] The manner of obtaining the hardware resource data includes: collecting classroom equipment operation state data and laboratory instrument use data through a campus Internet of Things sensor, exporting library collection book borrowing and inventory data through a library management system, and extracting campus network equipment load data through a network management platform. The manner of obtaining the software resource data includes: obtaining account active data and course resource access data from a teaching platform backend, obtaining user permission allocation data from a database management system, and extracting function use frequency data from a campus management system log. The manner of obtaining the teacher and student use demand data includes: collecting course and equipment reservation application data submitted by teachers and students through a campus mobile terminal, recording literature borrowing request data through a library borrowing system, and collecting bandwidth use demand data through a network traffic monitoring tool. The manner of obtaining the campus management demand data includes: extracting resource maintenance plan data from a campus logistics department file, obtaining energy consumption control index data from an energy consumption monitoring system, and extracting emergency resource dispatch demand data from an emergency management system.

[0023] Specifically, in the hardware resource data acquisition, the campus Internet of Things sensor (such as temperature sensor, current sensor) collects classroom equipment operation parameters (such as classroom intelligent eye protection lamp, intelligent blackboard lamp, equipment power consumption) and laboratory instrument use state (such as whether it is running, use time length) in real time, and the data sampling frequency is set to 1 time / minute to ensure real-time data; the library management system queries and exports the borrowing records (including borrower, borrowing time, return time) and inventory data (such as the number of books in the library, interlibrary loan situation) of the library collection books through SQL statements; the network management platform (such as Huawei eSight, H3CiMC) extracts port traffic, CPU utilization, bandwidth occupancy and other load data of campus network equipment through SNMP protocol, and the sampling period is set to 5 minutes, which takes into account real-time and data storage cost.

[0024] In the software resource data acquisition, the teaching platform backend obtains account active data (such as daily active user number, single active duration) and course resource access data (such as course click volume, video watching completion rate) through Python script, and the statistical period is 1 day; the database management system (such as MySQL, Oracle) obtains user access permission allocation data (such as accessible database list, permission validity period) of the academic database through permission query statement (GRANT / REVOKE); the campus management system log extracts use frequency data (such as daily use times, use user group) of each function module (such as repair module, procurement module) through log analysis tool (such as ELKStack).

[0025] In the process of obtaining the data of teachers and students' needs, the campus mobile terminal collects the data of course and equipment reservation application of teachers and students (including reservation subject, reservation resource, reservation period, and demand description) through the form submission function, and synchronizes the data to the background database in real time after submission. The library borrowing system records the literature borrowing request data (such as the name of the required literature, author, ISBN, and borrowing period) through the self-service terminal or online borrowing platform. The network traffic monitoring tool (such as Wireshark and NetflowAnalyzer) collects the network bandwidth usage demand data (such as peak bandwidth, average bandwidth, and bandwidth usage period) of each teaching building and dormitory area, with a precision of Mbps.

[0026] In the process of obtaining the data of campus management needs, the archives of campus logistics department are converted into structured data through document scanning and OCR recognition, and the resource maintenance plan data (such as maintenance equipment number, maintenance time, and maintenance personnel) are extracted. The energy consumption monitoring system (such as Schneider EcoStruxure) collects the regional energy consumption data through smart meters and water meters, and generates energy consumption control index data (such as monthly energy consumption upper limit and unit area energy consumption standard) combined with the campus management goals. The space management platform calculates the space utilization rate data by counting the usage frequency and duration of each area (such as classrooms, laboratories, and conference rooms) through video monitoring and infrared sensors. The emergency management system exports the emergency resource scheduling demand data (such as the number of fire extinguishers needed for fire scene and the type of first aid equipment needed for medical emergency) through emergency plan documents and historical emergency records.

[0027] The construction of the big data resource allocation analysis model includes: Determining the characteristic dimensions of campus resources and needs, the resource characteristic dimensions include resource type, available period, capacity specification, and maintenance cost, and the demand characteristic dimensions include demand subject, demand period, demand size, and emergency degree; Building a feature extraction module, converting resource-related data and demand-related data into standardized feature vectors through data segmentation, feature encoding, and outlier filtering processing; Building a demand matching module, setting matching rules, the matching rules include the coincidence degree rule of resource available period and demand period, the adaptation degree rule of resource capacity and demand size, and the corresponding rule of resource type and demand type; Building a deployment strategy generation module, importing campus resource allocation historical data and optimal case data, and training the strategy generation model through machine learning algorithm.

[0028] Specifically, the characteristic dimensions are determined as follows: In the resource characteristic dimension, resource types are classified according to "teaching / research / management / living" (e.g., classrooms belong to teaching, laboratories belong to research), available time periods are divided into "fixed / flexible / emergency call" (e.g., class time is fixed time period, after-class time is flexible time period), capacity specifications are defined according to the maximum usage that the resource can support (e.g., classroom capacity is 50 people, network bandwidth capacity is 1000Mbps), and maintenance costs are calculated according to the manpower and material costs required for a single maintenance (e.g., instrument maintenance cost is 500 yuan / time, network equipment maintenance cost is 200 yuan / time); In the demand characteristic dimension, the demand subjects are classified according to "students / teachers / administrators / visitors", demand time periods are precisely marked as "specific date + time period" (e.g., September 1, 2025, 9:00-11:00), demand scale is defined according to the number of people or the amount of resources involved in the demand (e.g., 50-person course demand, 100Mbps bandwidth demand), and urgency level is divided into "Level 1 / Level 2 / routine" (e.g., fire emergency is Level 1, large-scale event is Level 2, daily office is routine).

[0029] Feature extraction module construction: Data segmentation uses the Jieba segmentation tool to segment unstructured data (such as requirement description text) into words (e.g., "request to use a classroom for 50 people to take a math class" is segmented into "request / use / 50 people / classroom / take / math class"); Feature encoding uses One-Hot encoding to process categorical features (e.g., the resource type "teaching" is encoded as [1,0,0,0]) and Min-Max standardization to process numerical features (e.g., maintenance cost is converted from "500 yuan" to a standardized value in the range [0,1]); Outlier filtering uses the 3σ principle to identify and remove data that exceeds "mean ± 3 times standard deviation" (e.g., the maintenance cost of a certain device is 10,000 yuan, which is far beyond the mean of 500 yuan, and is identified as an outlier). Finally, the processed data is converted into a standardized feature vector with uniform dimensions (e.g., the resource feature vector is [1,0.8,0.5,0.3], and the requirement feature vector is [0.2,0.8,0.6,1]).

[0030] Detailed explanation of matching rules: The overlap rule between available resource periods and demand periods is as follows: Calculate the ratio of the overlap time between available resource periods and demand periods to the total demand period. A ratio of ≥80% is considered "high overlap", 60%-80% is "medium overlap", and <60% is "low overlap". The higher the overlap, the better the matching foundation. For example, if the available resource period is 9:00-12:00 and the demand period is 10:00-11:00, the overlap time is 1 hour, and the total demand time is 1 hour, the overlap is 100%, which is considered high overlap.

[0031] The matching rule between resource capacity and demand scale: Calculate the ratio of resource capacity to demand scale. A ratio between 1.0 and 1.2 is considered "matched" (resources are slightly greater than demand to avoid waste), 0.8 to 1.0 is considered "basically matched" (resources are slightly less than demand to meet core needs), and <0.8 or >1.2 is considered "not matched". For example, if the resource capacity is 50 people and the demand scale is 45 people, the ratio is 1.11, which is considered matched.

[0032] The rules for matching resource types with demand types are as follows: Establish a mapping table of "resource type - demand type" (e.g., teaching resources correspond to teaching needs, and research resources correspond to research needs). If the resource type and demand type directly correspond in the mapping table, it is judged as "completely correspondent"; if there is an indirect relationship (e.g., management resources can temporarily support teaching needs), it is judged as "partially correspondent"; if there is no relationship, it is judged as "not correspondent". Complete correspondence is the core premise of matching.

[0033] The resource allocation strategy generation module is built by importing historical data on campus resource allocation over the past three years (including historical allocation plans, execution results, and student and faculty feedback) and best-practice case data in the industry (such as successful cases of resource allocation in smart campuses of other universities). A strategy generation model is constructed using decision tree, random forest, and gradient boosting algorithms. During model training, the objective functions are "demand fulfillment rate, resource utilization rate, and student and faculty satisfaction." By adjusting algorithm parameters (such as decision tree depth and the number of random forest trees), the allocation strategy output by the model is made to match the objective function to the highest degree, ultimately forming a strategy generation model that can autonomously output initial allocation plans.

[0034] The resource types include teaching resources, research resources, management resources, and daily life resources; The available time periods include fixed time periods, flexible time periods, and emergency call time periods; The main groups demanding this information include students, teachers, administrative staff, and campus visitors. The urgency levels are divided into Level 1 Emergency, Level 2 Emergency, and Routine Needs. Level 1 Emergency corresponds to sudden emergency scenarios, Level 2 Emergency corresponds to short-term important activities, and Routine Needs corresponds to daily teaching and office needs.

[0035] Specifically, among the resource types, teaching resources refer to resources that directly serve teaching activities (such as classrooms, teaching projectors, and online course platforms); research resources refer to resources that support research projects (such as laboratory instruments, academic databases, and research funding); management resources refer to resources that ensure campus management (such as administrative offices, management system accounts, and logistical equipment); and living resources refer to resources that meet the living needs of teachers and students (such as canteen seats, dormitory water and electricity, and campus supermarket shelves). The classification is determined based on the "core purpose of the resource" to ensure coverage of all resource scenarios on campus.

[0036] In the available time slots, fixed time slots refer to the time slots where resources are pre-occupied and cannot be adjusted (such as classroom teaching time from 8:00 to 12:00 on weekdays), flexible time slots refer to the time slots where resources can be flexibly adjusted according to demand (such as laboratory time slots from 14:00 to 18:00 on weekdays), and emergency call time slots refer to the time slots where resources can only be used temporarily in emergency scenarios (such as the time slots for emergency equipment on standby 24 hours a day). The time slot division combines the daily operation rules of the campus with emergency needs to balance the stability and flexibility of resources.

[0037] Among the main groups with demand, students are further subdivided by grade (e.g., undergraduates, graduate students) or major (e.g., computer science students, humanities students); teachers are subdivided by teaching type (e.g., theoretical course teachers, experimental course teachers) or research direction; administrative staff are subdivided by department (e.g., academic affairs office, logistics office, finance office); and campus visitors are subdivided by purpose of visit (e.g., academic exchange, parent meeting, visit). This subdivision allows for a more precise matching of the needs of different groups (e.g., graduate students have higher requirements for the precision of research instruments than undergraduates).

[0038] In terms of urgency, Level 1 emergency needs have a response time of ≤1 hour (such as fire extinguisher dispatch during a fire or emergency equipment allocation during a sudden illness), and all available resources must be used first. Level 2 emergency needs have a response time of ≤24 hours (such as venue allocation for a campus recruitment fair in 3 days or equipment preparation for an academic conference in 1 week), and must be met first without affecting Level 1 needs. Routine needs have a response time of ≤7 days (such as classroom reservations for daily classes or borrowing ordinary documents), and are handled according to normal allocation procedures. The urgency level is determined based on the "scope of impact of the need and the response time requirement" to ensure that resources are prioritized for high-value needs.

[0039] The process of setting the matching rules includes: Collect successful and unsuccessful cases of resource allocation in the campus in the past, and extract the resource characteristics and demand characteristics parameters from the cases; The correlation coefficient between resource characteristics and demand characteristics in successful cases is calculated, and the correlation coefficient is obtained by Pearson correlation analysis and cosine similarity. The weight of each matching rule is determined based on the correlation coefficient. The higher the correlation coefficient, the greater the weight of the corresponding matching rule, thus forming the final multi-dimensional matching rule system.

[0040] Specifically, the process begins with case collection and parameter extraction. This involves collecting resource allocation cases from the past five years on campus. Successful cases must meet criteria such as "demand fulfillment rate ≥ 90%, resource utilization rate ≥ 85%, and student / faculty satisfaction ≥ 80%." For example, in the allocation of classrooms for final exams during a semester, all exam classes were precisely matched to classrooms of the corresponding specifications, with no resource waste or missed needs. Failed cases, on the other hand, fail to meet any of these criteria, such as a research team experiencing delays due to incorrect allocation of laboratory equipment. From each case, resource characteristic parameters (such as resource type and available time slots) and demand characteristic parameters (such as the demanding entity and demand time slot) are extracted and converted into calculable quantitative data. For instance, resource types are assigned values ​​such as "teaching = 1, research = 2," and urgency levels are assigned values ​​such as "routine = 1, secondary emergency = 2, primary emergency = 3," forming a case parameter matrix.

[0041] Next, the correlation coefficient is calculated using a combination of Pearson correlation analysis and cosine similarity. Pearson correlation analysis primarily examines the linear correlation between resource characteristics and demand characteristics. For example, when analyzing "available time periods" and "demand time periods," if their trends are highly consistent—for instance, if both resource availability and demand periods are predominantly daytime—their linear correlation is high, and the correlation coefficient is close to 1. If there is no clear pattern in their changes, the correlation coefficient is close to 0. Cosine similarity treats resource characteristics and demand characteristics as vectors and examines whether their directions are consistent. For example, with "resource capacity" and "demand scale," if a large resource capacity corresponds to a large demand scale, the two vectors are nearly aligned, resulting in a high similarity (close to 1). If the correspondence between resource capacity and demand scale is inconsistent, the similarity is low (close to 0). The results obtained from the two methods will then be standardized and averaged to obtain the comprehensive correlation coefficient. For example, the correlation coefficients for "available time period - demand time period" calculated by the two methods are 0.96 and 0.90 respectively, and the comprehensive correlation coefficient is 0.93; the comprehensive correlation coefficient for "resource type - demand type" is also 0.93; and the comprehensive correlation coefficient for "resource capacity - demand scale" is 0.9025.

[0042] Finally, the weights of the matching rules are determined, using the overall correlation coefficient of each rule as the basis for weight allocation. The higher the correlation coefficient, the greater the weight, and the sum of all rule weights is 1. For example, the sum of the overall correlation coefficients of the three matching rules mentioned earlier is 2.7625. Therefore, the weights of the "overlap of available time periods and demand time periods" rule and the "correspondence between resource type and demand type" rule are both approximately 0.336, while the weight of the "fitness between resource capacity and demand scale" rule is approximately 0.327. This forms a multi-dimensional matching rule system, allowing rules with high correlation to play a greater role in matching.

[0043] Inputting the resource-related data and demand-related data into the big data resource allocation and analysis model includes: The resource-related data and demand-related data are preprocessed, including data deduplication, missing value filling and data format conversion, to obtain preprocessed data. The preprocessed data is input into the feature extraction module to generate resource feature vectors and demand feature vectors. The resource feature vector and the demand feature vector are input into the demand matching module, and the matching score is calculated according to the preset matching rules. The matching score is obtained by weighted summation of each rule weight and the corresponding feature matching degree. Based on the matching scores, sort from high to low, and generate a list of campus resource supply and demand matching results.

[0044] Specifically, data preprocessing is performed first. During data deduplication, resource data is uniquely identified using "Resource ID + Data Collection Time," such as classroom equipment data using "Equipment Number + Collection Time," removing multiple records of the same equipment from the same time due to duplicate sensor reports. Demand data is uniquely identified using "Demand Submitter ID + Demand Submission Time + Demand Type," removing duplicate reservation or application data. When filling missing values, numerical data such as resource capacity is filled with the average of similar data; for example, if "Usage Time" of a laboratory instrument is missing, it is filled with the average usage time of similar instruments. Categorical data such as resource type is filled with the most frequently occurring category; for example, if "Demand Subject" of a demand is missing, it is filled with "Student Group," which appears most frequently in concurrent demands. For critical fields such as urgency, if the missing percentage exceeds 10%, the data is removed. During data format conversion, unstructured data, such as text describing the needs of teachers and students, will be converted into structured data by extracting keywords and then encoding them. For example, "application to use a high-precision oscilloscope for chip testing" will be converted into "high-precision oscilloscope, chip testing" by extracting keywords and encoding them as "scientific research instruments, scientific research needs". Time data in different formats will also be converted into the "YYYY-MM-DDHH:MM:SS" format.

[0045] Next, feature vectors are generated, and the preprocessed data is input into the feature extraction module. Different types of features are processed according to the feature dimensions defined in claim 4. Categorical features, such as resource type and demand subject, are encoded using One-Hot encoding; for example, "teaching resources" are encoded as [1,0,0,0], and "scientific research resources" are encoded as [0,1,0,0]. Numerical features, such as resource capacity and demand scale, are standardized, mapping the values ​​to the 0-1 range. For example, a resource capacity of "50 people" within the range of 20 to 100 people has a standardized value of 0.375. Temporal features, such as available time periods and demand periods, are first converted into timestamps and then standardized to the 0-1 range. Finally, fixed-dimensional resource feature vectors and demand feature vectors are generated. For example, the resource feature vector [1, 0.6, 0.375, 0.2] corresponds to "teaching resources, standardized value of available time period, standardized value of capacity specification, and standardized value of maintenance cost" respectively; the demand feature vector [0.2, 0.6, 0.35, 0.8] corresponds to "standardized value of student group, standardized value of demand time period, standardized value of demand scale, and standardized value of urgency level" respectively.

[0046] Then, the matching score is calculated based on the determined matching rules and weights. First, the matching degree of individual rule features is calculated. "Available Time Period - Demand Time Period Overlap" is the proportion of the overlap time between the two to the total demand time. For example, if the available time period of the resource is 0.2-0.6 (after standardization) and the demand time period is 0.3-0.5, the overlap time is 0.2, and the total demand time is 0.2, so the overlap degree is 1.0. "Resource Type - Demand Type Correspondence" is calculated as follows: if there is a complete correspondence, such as teaching resources matching teaching needs, the correspondence degree is 1.0; if there is a partial correspondence, such as management resources temporarily matching teaching needs, the correspondence degree is 0.6; and if there is no correspondence, the degree is 0. "Resource Capacity - Demand Scale Fit" is calculated based on the difference between the two standardized values. The smaller the difference, the higher the fit. For example, if the standardized value of the resource capacity is 0.375 and the standardized value of the demand scale is 0.35, the fit is 0.975. Next, the weighted matching score is calculated by multiplying the matching degree of each rule by its corresponding weight and then summing them. For example, if the weights of three rules are 0.336, 0.336, and 0.327, and their matching degrees are 1.0, 1.0, and 0.975, respectively, the weighted matching score will be approximately 0.9918. The score will then be limited to the range of 0-1, with a threshold of 0.6. "Resource-demand" pairs with a score ≥ 0.6 are considered valid matches and included in subsequent priority ranking; those with a score < 0.6 are excluded.

[0047] Finally, a list of matching results is generated, and the valid matching "resource-requirement" pairs are sorted from high to low according to the matching score. The list will contain "resource ID, requirement ID, matching score, resource characteristics, and requirement characteristics", such as resource A-requirement 1 (score 0.9918), resource B-requirement 2 (score 0.9752), etc., which provides a basis for determining the priority of resource allocation in the future.

[0048] Determining resource allocation priorities based on supply and demand matching results includes: Priority evaluation indicators are set, including the urgency of the need, resource utilization, allocation cost, and teacher and student satisfaction. Quantitative scoring standards were set for each evaluation indicator. The urgency of the demand was assigned different scores according to the urgency level. The resource utilization rate was calculated as the ratio of actual usage time to total available time. The allocation cost was calculated based on the costs of resource transportation, debugging and maintenance. Student and faculty satisfaction was obtained through questionnaires and user feedback scores. Calculate the overall priority score for each supply and demand matching item. The overall priority score is obtained by weighting the quantitative scores of each indicator with the preset indicator weights. The order of resource allocation is determined by sorting resources from highest to lowest based on their overall priority scores.

[0049] Specifically, priority evaluation indicators are first set, selecting four core indicators: "urgency of need, resource utilization rate, allocation cost, and student and faculty satisfaction," which respectively cover the dimensions of need value, resource efficiency, management cost, and user experience. Need urgency reflects the time sensitivity of the need; the higher the urgency, the higher the need value. Resource utilization reflects the efficiency of resource use; the lower the utilization rate, the more necessary it is to increase value through allocation. Allocation cost reflects the economic cost of allocation; the lower the cost, the higher the feasibility of allocation. Student and faculty satisfaction reflects the user acceptance of the solution; the higher the satisfaction, the more closely the solution meets the needs.

[0050] Further, a quantitative scoring standard is established. The urgency of the need is assigned a value according to the urgency level of claim 5. Level 1 urgency, such as a sudden emergency scenario, receives 100 points; Level 2 urgency, such as a short-term important event, receives 80 points; and routine needs, such as daily teaching and office work, receive 60 points. If the urgency level is upgraded, the value is reassigned according to the upgraded level. Resource utilization rate is calculated by dividing the actual usage time of the resource in the past 7 days by the total available time, and then standardizing the process to map the score to the range of 60-100. For example, if a classroom was actually used for 20 hours in the past 7 days and has a total available time of 35 hours, the utilization rate is approximately 57.14%. Within the range of resource utilization rates of 30%-80%, the standardized score is approximately 78 points. Allocation cost is the total cost of transporting, debugging, and maintaining resources, then assigned a value according to the cost range. For example, a cost below 500 yuan gets 100 points, 500-1000 yuan gets 80 points, 1000-2000 yuan gets 60 points, and above 2000 yuan gets 40 points. For example, if the total cost of transporting, debugging, and maintaining a batch of small teaching equipment is 300 yuan, it gets 100 points. Teacher and student satisfaction is obtained through questionnaires and user feedback. The average score of the questionnaires and feedback is taken and then mapped to a range of 60-100. For example, if the average score of teachers and students after the allocation of a certain resource is 85 points, it will be directly counted as 85 points.

[0051] Then, the overall priority score is calculated. First, each evaluation indicator is assigned a pre-defined weight, such as 0.4 for urgency, 0.2 for resource utilization, 0.2 for allocation cost, and 0.2 for student and faculty satisfaction, with a total weight of 1. Then, the quantitative score of each indicator for each supply-demand matching item is multiplied by its corresponding weight, and the results are summed to obtain the overall priority score. For example, if a matching item has 100 points for urgency, 78 points for resource utilization, 100 points for allocation cost, and 85 points for student and faculty satisfaction, the overall priority score would be 100×0.4+78×0.2+100×0.2+85×0.2=40+15.6+20+17=92.6 points. Finally, the allocation order was determined by sorting all supply and demand matching items from highest to lowest according to their comprehensive priority scores. Items with higher scores were allocated first, for example, matching items with a comprehensive score of 92.6 were ranked first and resources were allocated in advance. Items with lower scores were ranked in descending order to ensure that high-value, high-efficiency, low-cost, and high-satisfaction needs were met first.

[0052] Dynamically adjusting the initial resource allocation plan based on the execution data includes: Set the execution data collection cycle, which includes real-time collection, hourly collection and daily collection. Real-time collection is used for emergency resource allocation scenarios, while daily collection is used for regular teaching resource allocation scenarios. Extract key metrics from the execution data, including actual resource usage time, demand fulfillment rate, allocation delay time, and resource failure rate. The key indicators are compared with preset thresholds. If the key indicators exceed the preset threshold range, the reasons for the deviation are analyzed. The reasons for the deviation include excessive resource allocation, deviation in demand forecasting, and sudden equipment failure. Adjust the parameters of the big data resource allocation analysis model based on the cause of the deviation, regenerate the resource allocation plan, and achieve dynamic optimization.

[0053] Specifically, the data collection cycle is first set, and an appropriate collection frequency is selected based on different resource allocation scenarios. Emergency resource allocation scenarios, such as fire extinguisher dispatch during a fire or emergency equipment allocation during a sudden illness, require real-time monitoring of the execution status, so real-time collection is used, with a data collection interval of no more than 1 minute to ensure that problems can be detected immediately. Hourly collection is suitable for scenarios with certain timeliness requirements but no need for real-time monitoring, such as network bandwidth allocation during large-scale campus events, where data is collected hourly to understand bandwidth usage promptly without causing data redundancy. Regular teaching resource allocation scenarios, such as classroom allocation for daily classes, have relatively lower timeliness requirements, so daily collection is used, collecting data at a fixed time each day, such as 10 PM, which is sufficient to support subsequent adjustments and analysis.

[0054] Next, key metrics are extracted from the execution data. Actual resource usage time is the actual time a resource is used during the allocation period. For example, the actual time a classroom allocated for a course is used for class is the actual resource usage time. Demand fulfillment rate is the percentage of actually fulfilled demands out of the total demand. For example, if there are 10 classrooms needed during a certain period and 9 are actually fulfilled, the demand fulfillment rate is 90%. Allocation delay time is the time from when a demand is made to when the resource is actually allocated. For example, if a demand for laboratory equipment is made at 8:00 AM but is only allocated at 10:00 AM, the allocation delay time is 2 hours. Resource failure rate is the percentage of times a resource fails after allocation out of the total number of allocations. For example, if 1 out of 10 computers allocated fails, the resource failure rate is 10%.

[0055] Then, key indicators are compared with preset thresholds, and the reasons for deviations are analyzed. These preset thresholds are set based on the actual campus operations and management objectives. For example, the threshold for actual resource usage time is set as "no less than 70% of the allocated time." If a classroom is allocated for 8 hours but only used for 4 hours, falling below the threshold, the reasons for the deviation must be analyzed. Similarly, the threshold for demand fulfillment rate is no less than 90%. If the demand fulfillment rate is only 80% during a certain period, exceeding the threshold range, the reasons also need to be analyzed. Common reasons for deviations include: over-allocation of resources, such as allocating a classroom with a capacity of 50 people for a course, but only 20 students actually attending, resulting in wasted resources; demand forecasting errors, such as predicting 10 labs for a certain period, but only 5 actually being used, causing idle resources; and sudden equipment failures, such as allocated experimental instruments suddenly malfunctioning before use, failing to meet the demand.

[0056] Finally, the model parameters are adjusted and the plan is regenerated. The corresponding parameters of the big data resource allocation analysis model are adjusted according to different causes of deviation. If the deviation is due to excessive resource allocation, the parameters related to matching resource capacity with demand scale in the model are adjusted to make the resource capacity more closely match actual needs. If the deviation is due to demand forecasting errors, the parameters related to demand forecasting in the model are adjusted to improve forecasting accuracy. If the deviation is due to sudden equipment failure, the parameters related to resource failure early warning in the model are adjusted to mitigate failure risks in advance. After parameter adjustment, the latest resource data and demand data are input into the model to regenerate an optimized resource allocation plan, achieving dynamic optimization.

[0057] The process of setting the preset threshold includes: Collect data on campus resource allocation over the past three years and statistically analyze the historical averages and fluctuation ranges of key indicators. Based on campus management goals and the needs of teachers and students, the historical averages were revised. The threshold for actual resource usage time was set with reference to the standards for course duration and experiment duration. The threshold for demand satisfaction rate was no less than 90%, the threshold for allocation delay time was no more than 30 minutes, and the threshold for resource failure rate was no more than 5%. The revised thresholds were verified through an expert review meeting and adjusted in conjunction with the actual campus operation scenario to finally determine the preset threshold range for each key indicator.

[0058] Specifically, the first step is data collection and historical statistics. This involves comprehensively collecting execution data from all resource allocation scenarios across the campus over the past three years, covering different types of resources such as teaching resources (e.g., classrooms, teaching equipment), research resources (e.g., laboratory equipment, academic databases), and management resources (e.g., administrative office equipment, logistical tools), ensuring data coverage of core campus operations. For the four key indicators mentioned (actual resource usage time, demand fulfillment rate, allocation delay time, and resource failure rate), historical averages and fluctuation ranges are calculated. For example, the "actual resource usage time" for all classroom allocations over the past three years has an average of 75% of the total allocation time, with a fluctuation range of 60%-90%; the historical average for "demand fulfillment rate" is 88%, with a fluctuation range of 75%-95%; the historical average for "allocation delay time" is 25 minutes, with a fluctuation range of 10-40 minutes; and the historical average for "resource failure rate" is 6%, with a fluctuation range of 3%-10%.

[0059] Next, threshold adjustments are made based on management goals and demand standards. The core principle is to ensure that the thresholds both align with historical data patterns and meet the actual needs of campus management and faculty / students. For example, the "actual resource usage time threshold" references the correspondence between regular campus course durations (e.g., 45 minutes per theory class, 90 minutes per lab class) and resource allocation time. If a lab's allocation time is 90 minutes (matching one lab class), the actual usage time must at least cover the core components of the lab class (approximately 60 minutes). Based on the historical average of 75%, the threshold is adjusted to "no less than 70% of the allocation time" (e.g., 90 minutes of allocation time corresponds to an actual usage time of no less than 63 minutes). The "demand fulfillment rate threshold," considering the core demands of faculty / students regarding resource usage, sets the campus management goal at "regular..." The threshold for "meeting the demand in the specified scenarios should be no less than 90%" is revised to "no less than 90%" based on the historical average of 88%. The threshold for "allocation delay duration" is revised to "no more than 30 minutes" based on the time sensitivity of teachers and students to resources (e.g., ensuring classrooms are in place 30 minutes before class and instruments are debugged 30 minutes before experiments) and the historical average of 25 minutes. The threshold for "resource failure rate" is revised to "no more than 5%" based on the historical average of 6% to reduce delays caused by equipment failures, as the campus management goal is to keep the failure rate below 5%.

[0060] Finally, expert review and scenario adjustment will be conducted. A review meeting composed of personnel from campus logistics management, teaching management, research departments, and big data technology experts will be organized to verify the revised thresholds from multiple dimensions. On the one hand, the "feasibility" of the thresholds will be verified. For example, regarding "resource failure rate not exceeding 5%", experts will consider the aging of existing campus equipment and maintenance frequency to determine whether this threshold can be achieved by strengthening daily maintenance. On the other hand, the "reasonableness" of the thresholds will be verified. For example, regarding "allocation delay time not exceeding 30 minutes", experts will consider the distribution of campus resources (such as the distance between teaching buildings and laboratory buildings, and equipment transportation time) to determine whether this threshold conforms to the actual operation scenario. If problems are found during the review, targeted adjustments will be made. For example, for a remote campus where the 30-minute allocation delay threshold is difficult to achieve due to the long resource transportation distance, after expert discussion, the "allocation delay time threshold" for that campus will be adjusted to "not exceeding 45 minutes", while retaining the 30-minute threshold for other core campuses. Another example is for precision scientific research instruments, where the maintenance difficulty is high, the threshold of "resource failure rate not exceeding 5%" needs to be lowered to "not exceeding 3%" to meet the stability requirements of scientific research experiments. Ultimately, a "differentiated preset threshold system" covering different resource types and different campus scenarios was formed to ensure that each threshold can be implemented.

[0061] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for dynamic optimization and allocation of smart campus resources based on big data, characterized in that, Includes the following steps: Acquire resource-related data and demand-related data for smart campuses. The resource-related data includes campus hardware resource data and software resource data, and the demand-related data includes data on the usage needs of teachers and students and data on campus management needs. A big data resource allocation and analysis model is constructed, which includes a resource feature extraction module, a demand matching module, and an allocation strategy generation module. The resource-related data and demand-related data are input into the big data resource allocation and analysis model to obtain the campus resource supply and demand matching results. Determine resource allocation priorities based on supply and demand matching results, and generate an initial resource allocation plan based on the allocation priorities; The execution data during the resource allocation process is collected in real time, and the initial resource allocation plan is dynamically adjusted based on the execution data to obtain an optimized campus resource allocation plan.

2. The method for dynamic optimization and allocation of smart campus resources based on big data according to claim 1, characterized in that, The hardware resource data includes classroom equipment data, laboratory equipment data, library collection data, and campus network equipment data; The software resource data includes teaching platform account data, academic database permission data, campus management system function data, and online course resource data. The data on the needs of teachers and students includes data on course reservation needs, data on experimental equipment usage needs, data on document borrowing needs, and data on network bandwidth needs. The campus management needs data include resource maintenance plan data, energy consumption control needs data, space utilization improvement needs data, and emergency resource dispatch needs data.

3. The method for dynamic optimization and allocation of smart campus resources based on big data according to claim 2, characterized in that, The methods for obtaining the hardware resource data include: collecting classroom equipment operation status data and laboratory instrument usage data through campus IoT sensors, exporting library book borrowing and inventory data through the library management system, and extracting campus network equipment load data through the network management platform; The methods for obtaining the software resource data include: collecting account activity data and course resource access data from the teaching platform backend, obtaining user permission allocation data from the database management system, and extracting function usage frequency data from the campus management system logs; The methods for obtaining the data on the usage needs of teachers and students include: collecting course and equipment reservation application data submitted by teachers and students through the campus mobile terminal, recording document borrowing request data through the library borrowing system, and collecting bandwidth usage demand data through network traffic monitoring tools; The methods for obtaining the campus management demand data include: extracting resource maintenance plan data from the archives of the campus logistics department, obtaining energy consumption control index data from the energy consumption monitoring system, compiling utilization rate data for each area from the space management platform, and exporting emergency resource dispatch demand data from the emergency management system.

4. The method for dynamic optimization and allocation of smart campus resources based on big data according to claim 1, characterized in that, The construction of the big data resource allocation and analysis model includes: Determine the dimensions of campus resource characteristics and demand characteristics. The resource characteristics include resource type, available time period, capacity specifications and maintenance costs. The demand characteristics include demand subject, demand time period, demand scale and urgency. A feature extraction module was built, which converts resource-related data and demand-related data into standardized feature vectors through data segmentation, feature encoding, and outlier filtering. Build a demand matching module and set matching rules, including rules for the overlap between resource availability time periods and demand time periods, rules for the fit between resource capacity and demand scale, and rules for the correspondence between resource type and demand type. A strategy generation module was built, which imported historical data and best case data on campus resource allocation, and trained the strategy generation model through machine learning algorithms.

5. The method for dynamic optimization and allocation of smart campus resources based on big data according to claim 4, characterized in that, The resource types include teaching resources, research resources, management resources, and daily life resources; The available time periods include fixed time periods, flexible time periods, and emergency call time periods; The main groups demanding this information include students, teachers, administrative staff, and campus visitors. The urgency levels are divided into Level 1 Emergency, Level 2 Emergency, and Routine Needs. Level 1 Emergency corresponds to sudden emergency scenarios, Level 2 Emergency corresponds to short-term important activities, and Routine Needs corresponds to daily teaching and office needs.

6. The method for dynamic optimization and allocation of smart campus resources based on big data according to claim 4, characterized in that, The process of setting the matching rules includes: Collect successful and unsuccessful cases of resource allocation in the campus in the past, and extract the resource characteristics and demand characteristics parameters from the cases; The correlation coefficient between resource characteristics and demand characteristics in successful cases is calculated, and the correlation coefficient is obtained by Pearson correlation analysis and cosine similarity. The weight of each matching rule is determined based on the correlation coefficient. The higher the correlation coefficient, the greater the weight of the corresponding matching rule, thus forming the final multi-dimensional matching rule system.

7. The method for dynamic optimization and allocation of smart campus resources based on big data according to claim 1, characterized in that, Inputting the resource-related data and demand-related data into the big data resource allocation and analysis model includes: The resource-related data and demand-related data are preprocessed, including data deduplication, missing value filling and data format conversion, to obtain preprocessed data. The preprocessed data is input into the feature extraction module to generate resource feature vectors and demand feature vectors. The resource feature vector and the demand feature vector are input into the demand matching module, and the matching score is calculated according to the preset matching rules. The matching score is obtained by weighted summation of each rule weight and the corresponding feature matching degree. Based on the matching scores, sort from high to low, and generate a list of campus resource supply and demand matching results.

8. The method for dynamic optimization and allocation of smart campus resources based on big data according to claim 1, characterized in that, Determining resource allocation priorities based on supply and demand matching results includes: Priority evaluation indicators are set, including the urgency of the need, resource utilization rate, allocation cost, and teacher and student satisfaction. Quantitative scoring standards were set for each evaluation indicator. The urgency of the demand was assigned different scores according to the urgency level. The resource utilization rate was calculated as the ratio of actual usage time to total available time. The allocation cost was calculated based on the costs of resource transportation, debugging and maintenance. Student and faculty satisfaction was obtained through questionnaires and user feedback scores. Calculate the overall priority score for each supply and demand matching item. The overall priority score is obtained by weighting the quantitative scores of each indicator with the preset indicator weights. The order of resource allocation is determined by sorting resources from highest to lowest based on their overall priority scores.

9. The method for dynamic optimization and allocation of smart campus resources based on big data according to claim 1, characterized in that, Dynamically adjusting the initial resource allocation plan based on the execution data includes: Set the execution data collection cycle, which includes real-time collection, hourly collection and daily collection. Real-time collection is used for emergency resource allocation scenarios, while daily collection is used for regular teaching resource allocation scenarios. Extract key metrics from the execution data, including actual resource usage time, demand fulfillment rate, allocation delay time, and resource failure rate. The key indicators are compared with preset thresholds. If the key indicators exceed the preset threshold range, the reasons for the deviation are analyzed. The reasons for the deviation include excessive resource allocation, deviation in demand forecasting, and sudden equipment failure. Adjust the parameters of the big data resource allocation analysis model based on the cause of the deviation, regenerate the resource allocation plan, and achieve dynamic optimization.

10. A method for dynamic optimization and allocation of smart campus resources based on big data as described in claim 9, characterized in that, The process of setting the preset threshold includes: Collect data on campus resource allocation over the past three years and statistically analyze the historical averages and fluctuation ranges of key indicators. Based on campus management goals and the needs of teachers and students, the historical averages were revised. The threshold for actual resource usage time was set with reference to the standards for course duration and experiment duration. The threshold for demand satisfaction rate was no less than 90%, the threshold for allocation delay time was no more than 30 minutes, and the threshold for resource failure rate was no more than 5%. The revised thresholds were verified through an expert review meeting and adjusted in accordance with the actual operation of the campus, ultimately determining the preset threshold range for each key indicator.

Images