A temperature and humidity control method based on improved genetic algorithm

Abstract

The application discloses a temperature and humidity control method based on an improved genetic algorithm, and relates to the technical field of data processing.The method comprises the following steps: collecting temperature and humidity data of an environment where a server is located; performing analog-digital conversion on the temperature and humidity data by using an Arduino, and transmitting the temperature and humidity data to a Raspberry Pi; on one hand, the Raspberry Pi collects the temperature and humidity data, processes the data by using the improved genetic algorithm, determines an optimal solution of a temperature control device and a dehumidification device, and performs adaptive temperature and humidity control, wherein the improved genetic algorithm is obtained by introducing a gene cluster after improving a fitness function, a selection operator, a crossover operator and a mutation operator of a standard genetic algorithm; on the other hand, the Raspberry Pi serves as a gateway, uploads the processed data to a Tsinghua Tongjiang cloud server for persistence; a user logs in to the Tsinghua Tongjiang cloud server through a Web webpage, remotely checks real-time temperature and humidity data and historical temperature and humidity data of the environment where the server is located, and remotely controls temperature and humidity devices.The application can reduce the computing power pressure of the cloud server by collecting and processing data through the Raspberry Pi.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a temperature and humidity control method based on improved genetic algorithm. BACKGROUND

[0002] At present, the traditional PID algorithm is usually used for adaptive control of temperature and humidity of server room, and in this process, the cloud server needs to bear huge computing power pressure.

[0003] Therefore, the researchers found that the edge computing capability of Raspberry Pi can be used to quickly find the optimal temperature and humidity of the server room by adjusting the search direction adaptively through the genetic algorithm, thereby reducing the computing power pressure of the cloud server.

[0004] Genetic algorithm (GA) is a method for searching optimal solution by simulating natural evolution process, and its main features are directly operating on the structure object, without derivation and function continuity limitation, and without the need of determining rules to automatically obtain and guide the optimization search space and adaptively adjust the search direction.

[0005] Raspberry Pi is a RAM-based microcomputer motherboard, which uses SD / MicroSD card as internal hard disk, and has 1 / 2 / 4 USB interfaces and a 10 / 100 Ethernet interface around the card motherboard, which can be connected with keyboard, mouse and network cable, and has TV output interface and HDMI high-definition video output interface of video analog signal, all of which are integrated on a motherboard only slightly larger than a credit card, small in size, low in power consumption and rich in functions, and very suitable for embedded home gateway.

[0006] The wave cloud server provides storage of massive device data, opens rich device management interfaces, saves the cost of managing data, and provides the ability of remote control, remote debugging, remote upgrading and remote configuration of devices, so that the device can be managed without going to the device site. SUMMARY

[0007] The present application provides a temperature and humidity control method based on improved genetic algorithm to meet the needs and deficiencies of current technology development.

[0008] The technical solution for solving the above technical problems of the temperature and humidity control method based on improved genetic algorithm is as follows:

[0009] A temperature and humidity control method based on improved genetic algorithm, comprising the following steps:

[0010] Collecting temperature and humidity data of the environment where the server is located;

[0011] Arduino converts the temperature and humidity data into analog signals and transmits the analog signals to the edge service node Raspberry Pi through serial communication;

[0012] The edge service node Raspberry Pi aggregates the temperature and humidity data, processes the data by using the improved genetic algorithm, determines the optimal solution of the temperature control device and the dehumidification device, and controls the start and stop of the temperature control device and the dehumidification device to control the temperature and humidity, wherein the improved genetic algorithm is obtained by improving the fitness function, the selection operator, the crossover operator and the mutation operator of the standard genetic algorithm and introducing the gene cluster, and the edge service node Raspberry Pi serves as a gateway to upload the processed data to the Inspur cloud server for persistence.

[0013] The user logs in to the Inspur cloud server through a Web page, remotely views the real-time temperature and humidity data and the historical temperature and humidity data of the environment where the server is located, and manually controls the temperature control device and the dehumidification device remotely.

[0014] Optionally, the improvement on the fitness function of the standard genetic algorithm is as follows:

[0015] A fitness function that can dynamically change the scale of the fitness value according to the evolution process of the population is designed, as shown in formula (1)

[0016]

[0017] wherein f is the fitness value of a certain individual, f min is the minimum fitness value of the individuals in the current population, f max is the maximum fitness value of the individuals in the current population, and a is a constant.

[0018] Further optionally, the improvement on the selection operator of the standard genetic algorithm is as follows: when selecting individuals, a selection strategy of multiple copies of excellent individuals without replacement is adopted, and the individuals whose fitness values exceed a set threshold are retained, and the process specifically includes

[0019] (2.1) setting the number n of individuals to be selected;

[0020] (2.2) selecting n individuals without replacement according to the roulette selection algorithm;

[0021] (2.3) sorting the n individuals according to the fitness values from large to small, and copying the first 1 / 3 excellent individuals twice and the middle 1 / 3 individuals once;

[0022] (2.4) after the parent mating population is crossed and mutated, the individuals whose fitness values exceed a set threshold are retained.

[0023] Further, the improvement of the crossover operator and mutation operator of the standard genetic algorithm is that: adaptive groups of the crossover operator and mutation operator are preconfigured, so that different adaptive groups of the crossover operator and mutation operator are dynamically selected according to different stages of population evolution, which is specifically that: in the initial stage of population evolution, the crossover operator and mutation operator in the adaptive group which can realize strong crossover and weak mutation are selected, when the population evolution is slow and falls into a local optimal solution, the crossover operator and mutation operator in the adaptive group which can realize weak crossover and strong mutation are selected;

[0024] Before the adaptive groups of the crossover operator and mutation operator are preconfigured, the following calculation is needed:

[0025] The difference △f between the maximum fitness value fmax of the individual in the current population and the average fitness value fave of all individuals in the current population is calculated by formula (2): max

[0026]

[0027] The fitness value is mapped to the 0-1 probability interval by formula (3):

[0028]

[0029] Wherein, σ is the sigmod function.

[0030] Further, the operation process of the improved genetic algorithm for determining the optimal solution is as follows:

[0031] (1) Encoding;

[0032] (2) Determine the fitness function;

[0033] (3) Initialize the population, calculate the individual fitness value, and calibrate the fitness value;

[0034] (4) When selecting individuals, the selection strategy of multiple copies of excellent individuals without replacement is adopted, and the individuals whose fitness values exceed the set threshold in the population are reserved;

[0035] (5) Preconfigure adaptive groups of the crossover operator and mutation operator, dynamically select the crossover operator and mutation operator of different adaptive groups according to different stages of population evolution, and perform crossover operation and mutation operation;

[0036] (6) The intergenerational subspace optimization gene cluster strategy is adopted to determine the optimal solution, and the specific implementation process is as follows:

[0037] ​​(6.1) Partitioning the optimization space: find all the continuous non-intersecting paths in the individual, and the ordered set of cities in each path, the number of sets is n, a set is called a gene cluster, then m individuals are divided into n gene clusters, and the m individuals are arranged in descending order according to the fitness value as formula (4), and the sorted individuals are placed in set S,

[0038] B1, B2,..., B i ,..., B m , f(Bi) >= f(B i+1 ), i = 1, 2,..., m (4);

[0039] (6.2) Optimize the gene cluster: select the top ηm individuals in the order, as shown in formula (5), ηm individuals are given a weight λ i according to the principle of high weight in the order, find the same gene cluster between individuals and record it, and the similar gene cluster between individuals is retained according to the product value of the fitness value g(.) of the gene cluster and the weight λ i ,

[0040] B i e S i , i = 1, 2,..., ηm (5),

[0041] Wherein, η is a probability artificially set;

[0042] (6.3) Parent cross, generate child gene cluster: each gene cluster is regarded as a gene, when crossing, two individuals exchange part of the gene cluster, and the genes other than the gene cluster are crossed freely, and after generating the child individual, find the new gene cluster;

[0043] (6.4) Improve the fitness of the gene cluster by continuous evolution, and when the fitness of the gene cluster reaches the set maximum threshold, randomly select an individual in the gene cluster as the optimal solution;

[0044] (7) Determine whether the optimal solution meets the preset termination condition, if yes, output the optimal solution, if not, return to step (3) to recalculate the fitness value of the individual.

[0045] Optionally, the temperature and humidity digital signals of the environment where the server is located are collected by the temperature and humidity sensor DHT11, and the temperature and humidity digital signals are processed by the Arduino to obtain intuitive temperature and humidity values, which are then transmitted to the edge service node Raspberry Pi through serial communication.

[0046] Further optionally, the temperature and humidity sensor DHT11 is installed in the machine room and the server air outlet.

[0047] Optionally, an ECharts component is deployed in the Inspur cloud server, and the Inspur cloud server displays the received data on a web page in the form of a line chart, a histogram, or a chart through the ECharts component.

[0048] Optionally, mosquito is installed on the Inspur cloud server and the edge service node Raspberry Pi, and the edge service node Raspberry Pi transmits data with the Inspur cloud server through the MQTT protocol.

[0049] Users log in to the Inspur cloud server through the web page and manually control the temperature control equipment and dehumidification equipment through the MQTT protocol;

[0050] The edge business node Raspberry Pi uses the MQTT protocol to adaptively control temperature control equipment and dehumidification equipment.

[0051] The temperature and humidity control method based on the improved genetic algorithm of the present invention has the following beneficial effects compared with the prior art:

[0052] (1) The present invention uses Raspberry Pi to aggregate temperature and humidity data. Raspberry Pi uses an improved genetic algorithm to process the data, determine the optimal solution for the temperature control device and the dehumidification device, and automatically adjust the temperature control device and the dehumidification device, thereby reducing the computing pressure of the cloud server;

[0053] (2) The present invention uploads data to the cloud server through Raspberry Pi, and displays the data through the ECharts component of the cloud server, making it convenient for users to view environmental data in real time; after the data is uploaded to the Inspur cloud server for persistence, the ECharts component can also draw historical temperature and humidity data into curves for easy viewing. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Attachment Figure 1 This is a structural hierarchy diagram of the temperature and humidity control method of the present invention;

[0055] Attachment Figure 2 It is the flow chart of the improved genetic algorithm of the present invention. DETAILED DESCRIPTION

[0056] In order to make the technical solution, the technical problems solved and the technical effects of the present invention more clear, the technical solution of the present invention is clearly and completely described below in conjunction with specific embodiments.

[0057] Example 1:

[0058] Combined with attachment Figure 1 This embodiment proposes a temperature and humidity control method based on an improved genetic algorithm, comprising the following steps:

[0059] (I) The machine room and the server air outlet are provided with a temperature and humidity sensor DHT11, which collects the temperature and humidity digital signals of the environment where the server is located.

[0060] (II) The Arduino performs analog-to-digital conversion on the temperature and humidity digital signals to obtain intuitive temperature and humidity values, which are then transmitted to the edge service node Raspberry Pi through serial communication.

[0061] (III) The edge service node Raspberry Pi, on one hand, aggregates temperature and humidity data, processes the data using an improved genetic algorithm to determine the optimal solution of the temperature control device and the dehumidification device, and controls the start and stop of the temperature control and dehumidification devices to perform temperature and humidity control edge service nodes. Raspberry Pi, on the other hand, serves as a gateway to upload the processed data to the Inspur cloud server for persistence.

[0062] Among them, the improved genetic algorithm is obtained by improving the fitness function, selection operator, crossover operator and mutation operator of the standard genetic algorithm and introducing gene clusters.

[0063] (i) The improvement of the fitness function of the standard genetic algorithm is as follows:

[0064] A fitness function that can dynamically change the scale of fitness value according to the evolution process of the population is designed, as shown in formula (1)

[0065]

[0066] Where f is the fitness value of a certain individual, f min is the minimum fitness value of the individuals in the current population, f max is the maximum fitness value of the individuals in the current population, and a is a constant.

[0067] (ii) The improvement of the selection operator of the standard genetic algorithm is as follows: when selecting individuals, a selection strategy of multiple copies of excellent individuals without replacement is adopted, and individuals with fitness values exceeding a certain threshold are retained. This process specifically includes

[0068] (2.1) Set the number of individuals to be selected n;

[0069] (2.2) Select n individuals without replacement according to the roulette selection algorithm;

[0070] (2.3) Sort the n individuals according to the fitness value from large to small, and copy the first 1 / 3 excellent individuals twice and the middle 1 / 3 individuals once;

[0071] (2.4) After crossover and mutation, the parent mating population retains individuals whose fitness values ​​exceed the set threshold.

[0072] (iii) The crossover and mutation operators of the standard genetic algorithm are improved by pre-configuring adaptive groups of crossover and mutation operators so that crossover and mutation operators from different adaptive groups can be dynamically selected according to the different stages of population evolution. Specifically, in the initial stage of population evolution, crossover and mutation operators from the adaptive group that can achieve strong crossover and weak mutation are selected. When the population evolves slowly and falls into a local optimal solution, crossover and mutation operators from the adaptive group that can achieve weak crossover and strong mutation are selected.

[0073] Before pre-configuring the adaptive groups of crossover and mutation operators, the following calculations need to be performed:

[0074] The maximum fitness value f of individuals in this generation population is calculated by formula (2) max The average fitness value of all individuals in the current generation population The difference △f,

[0075]

[0076] The fitness value is mapped to the 0-1 probability interval through formula (3),

[0077]

[0078] Where σ is the sigmoid function.

[0079] Combined with attachment Figure 2 ,The operation process of improving the genetic algorithm to determine the optimal solution is as follows:

[0080] (1) Coding;

[0081] (2) Determine the fitness function;

[0082] (3) Initialize the population, calculate the individual fitness value, and calibrate the fitness value;

[0083] (4) When selecting individuals, a selection strategy of multiple replications of excellent individuals without replacement is adopted to retain individuals in the population whose fitness values ​​exceed the set threshold;

[0084] (5) Pre-configure adaptive groups of crossover operators and mutation operators, and dynamically select crossover operators and mutation operators from different adaptive groups according to different stages of population evolution to perform crossover and mutation operations;

[0085] (6) Adopt the intergenerational subspace optimization gene cluster strategy to determine the optimal solution. The specific implementation process is as follows:

[0086] (6.1) Divide the optimization space: Find the ordered set of cities in all continuous uncrossed paths among individuals. The number of sets is n. A set is called a gene cluster. Then, m individuals are divided into n gene clusters. Arrange the m individuals in descending order according to the fitness value as formula (4). Put the sorted individuals into the set S.

[0087] B1, B2, ..., B i ,...,B m ,f(B i )≥f(B i+1 ), i=1, 2,..., m (4);

[0088] (6.2) Gene cluster optimization: Select the first ηm individuals after sorting, as shown in formula (5), and give a weight λ to each of the ηm individuals according to the principle that the first one has the highest weight. i , find the same gene clusters between individuals and record them, similar gene clusters between individuals, according to the fitness value g(.) of the gene cluster and the weight λ i The product value is greater than the set value, and the gene clusters are retained.

[0089] B i ∈S i , i=1, 2,..., ηm (5),

[0090] Among them, η is an artificially set probability;

[0091] (6.3) Parent crossover to generate offspring gene clusters: Each gene cluster is considered as a gene. During crossover, two individuals exchange part of their gene clusters. Genes outside the gene clusters are freely crossed over. After the offspring individuals are generated, new gene clusters are found.

[0092] (6.4) Through continuous evolution, the fitness of the gene cluster is improved. When the fitness of the gene cluster reaches the set maximum threshold, an individual in the gene cluster is randomly selected as the optimal solution;

[0093] (7) Determine whether the optimal solution meets the preset termination condition. If it does, output the optimal solution. If not, return to step (3) and recalculate the individual fitness value.

[0094] (4) Users log in to the Inspur cloud server through a web page, remotely view the real-time and historical temperature and humidity data of the environment where the server is located, and manually remotely control the temperature control equipment and dehumidification equipment.

[0095] The ECharts component is deployed in the Inspur Cloud Server. The Inspur Cloud Server displays the received data on the Web page in the form of a line chart, histogram or chart through the ECharts component.

[0096] In the embodiment, the Tsinghua cloud server and the edge service node Raspberry Pi are installed with mosquito, and the edge service node Raspberry Pi transmits data with the Tsinghua cloud server through the MQTT protocol.

[0097] The user logs in the Tsinghua cloud server through a Web page, and manually controls the temperature control equipment and the dehumidification equipment through the MQTT protocol.

[0098] The edge service node Raspberry Pi controls the temperature control equipment and the dehumidification equipment through the MQTT protocol.

[0099] As can be seen, the temperature and humidity control method based on the improved genetic algorithm can automatically adjust the temperature control equipment and the dehumidification equipment through Raspberry Pi, reduce the computing power pressure of the cloud server, and realize data display through the ECharts component of the cloud server, so that the user can conveniently view the environmental data in real time. The operation and maintenance personnel can obtain the computer room environmental information outside the computer room, and the method is convenient for people's life.

[0100] The above application specific examples have described the principles and implementation modes of the application in detail, and these examples are only used to help understand the core technical content of the application. Based on the above specific embodiments of the application, any improvement and modification of the application made by the person skilled in the art without departing from the principles of the application shall fall within the patent protection scope of the application.

Claims

1. A temperature and humidity control method based on an improved genetic algorithm, characterized in that , including the following steps: Collect temperature and humidity data of the server's environment; Arduino performs analog-to-digital conversion on the temperature and humidity data and transmits it to the edge service node RaspberryPi via serial communication; The edge service node Raspberry Pi aggregates temperature and humidity data, processes the data using an improved genetic algorithm, determines the optimal solution for temperature control and dehumidification equipment, and adaptively controls the start and stop of the temperature control and dehumidification equipment to control temperature and humidity. The improved genetic algorithm is an improvement on the fitness function, selection operator, crossover operator, and mutation operator of the standard genetic algorithm, and is obtained by introducing a gene cluster. The edge service node Raspberry Pi also acts as a gateway to upload the processed data to the Inspur cloud server for persistence. Users log in to the Inspur cloud server through the web page, remotely view the real-time and historical temperature and humidity data of the server's environment, and manually remotely control the temperature control and dehumidification equipment.

2. The temperature and humidity control method based on the improved genetic algorithm according to claim 1, characterized in that: The improvement to the fitness function of the standard genetic algorithm is: Design a fitness function that can dynamically change the scale of fitness value according to the population evolution process, as shown in formula (1) Among them, f is the fitness value of an individual, f min is the minimum fitness value of individuals in this generation population, f max It is the maximum fitness value of individuals in this generation population, and α is a constant.

3. The temperature and humidity control method based on the improved genetic algorithm according to claim 2, characterized in that: The improvement to the selection operator of the standard genetic algorithm is as follows: when selecting individuals, a selection strategy of multiple copies of excellent individuals without replacement is adopted to retain individuals in the population whose fitness values ​​exceed the set threshold. This process specifically includes (2.1) Set the number of individuals n to be selected; (2.2) According to the roulette wheel selection algorithm, n individuals are selected without replacement; (2.3) Sort n individuals by their fitness values ​​from large to small, and make two copies of the top 1 / 3 of the best individuals and one copy of the middle 1 / 3 of the individuals; (2.4) After crossover and mutation, the parent mating population retains individuals whose fitness values ​​exceed the set threshold.

4. The temperature and humidity control method based on the improved genetic algorithm according to claim 3, characterized in that: The crossover and mutation operators of the standard genetic algorithm are improved by pre-configuring adaptive groups of crossover and mutation operators, so that crossover and mutation operators from different adaptive groups can be dynamically selected according to the different stages of population evolution. Specifically, in the initial stage of population evolution, crossover and mutation operators from the adaptive group that can achieve strong crossover and weak mutation are selected. When the population evolves slowly and falls into a local optimal solution, crossover and mutation operators from the adaptive group that can achieve weak crossover and strong mutation are selected. Before pre-configuring the adaptive groups of crossover and mutation operators, the following calculations need to be performed: The maximum fitness value f of individuals in this generation population is calculated by formula (2) max The average fitness value of all individuals in the current generation population The difference △f, The fitness value is mapped to the 0-1 probability interval through formula (3), Where σ is the sigmoid function.

5. The temperature and humidity control method based on the improved genetic algorithm according to claim 4, characterized in that: The operation flow of the improved genetic algorithm to determine the optimal solution is as follows: (1) Coding; (2) Determine the fitness function; (3) Initialize the population, calculate the individual fitness value, and calibrate the fitness value; (4) When selecting individuals, a selection strategy of multiple replications of excellent individuals without replacement is adopted to retain individuals in the population whose fitness values ​​exceed the set threshold; (5) Pre-configure adaptive groups of crossover operators and mutation operators, and dynamically select crossover operators and mutation operators from different adaptive groups according to different stages of population evolution to perform crossover and mutation operations; (6) Adopting the intergenerational subspace optimal gene clustering strategy to determine the optimal solution; (7) Determine whether the optimal solution meets the preset termination condition. If it does, output the optimal solution. If not, return to step (3) and recalculate the individual fitness value.

6. The temperature and humidity control method based on the improved genetic algorithm according to claim 5, characterized in that: The specific implementation process of step (6) is as follows: (6.1) Divide the optimization space: Find the ordered set of cities in all continuous uncrossed paths among individuals. The number of sets is n. A set is called a gene cluster. Then, m individuals are divided into n gene clusters. Arrange the m individuals in descending order according to the fitness value as formula (4). Put the sorted individuals into the set S. B1,B2,...,B i ,...,B m ,f(B i )≥f(B i+1 ),i=1,2,...,m (4); (6.2) Gene cluster optimization: Select the first ηm individuals after sorting, as shown in formula (5), and give a weight λ to each of the ηm individuals according to the principle that the first one has the highest weight. i , find the same gene clusters between individuals and record them, similar gene clusters between individuals, according to the fitness value g(.) of the gene cluster and the weight λ i The product value is greater than the set value, and the gene clusters are retained. B i ∈S i ,i=1,2,...,ηm (5), Among them, η is an artificially set probability; (6.3) Parent crossover to generate offspring gene clusters: Each gene cluster is considered as a gene. During crossover, two individuals exchange part of their gene clusters. Genes outside the gene clusters are freely crossed over. After the offspring individuals are generated, new gene clusters are found. (6.4) Through continuous evolution, the fitness of the gene cluster is improved. When the fitness of the gene cluster reaches the set maximum threshold, an individual in the gene cluster is randomly selected as the optimal solution.

7. The temperature and humidity control method based on an improved genetic algorithm according to claim 1, characterized in that: The temperature and humidity digital signals of the server's environment are collected through the temperature and humidity sensor DHT11. Arduino processes the temperature and humidity digital signals to obtain intuitive temperature and humidity values, and then transmits them to the edge business node RaspberryPi through serial communication.

8. The temperature and humidity control method based on an improved genetic algorithm according to claim 7, characterized in that: The temperature and humidity sensor DHT11 is installed in the computer room and the server air outlet.

9. The temperature and humidity control method based on an improved genetic algorithm according to claim 1, characterized in that: The Inspur cloud server is deployed with an ECharts component, and the Inspur cloud server displays the received data on a web page in the form of a line chart, a histogram or a diagram through the ECharts component.

10. The temperature and humidity control method based on an improved genetic algorithm according to claim 1, characterized in that: Mosquito is installed on the Inspur cloud server and the edge service node Raspberry Pi; Users log in to the Inspur cloud server through the web page and manually control the temperature control equipment and dehumidification equipment through the MQTT protocol; The edge service node Raspberry Pi performs adaptive control of the temperature control device and the dehumidification device through the MQTT protocol.

Images