CN112464311A - Equipment arrangement method of smart city data center - Google Patents

Equipment arrangement method of smart city data center Download PDF

Info

Publication number
CN112464311A
CN112464311A CN202011332596.7A CN202011332596A CN112464311A CN 112464311 A CN112464311 A CN 112464311A CN 202011332596 A CN202011332596 A CN 202011332596A CN 112464311 A CN112464311 A CN 112464311A
Authority
CN
China
Prior art keywords
room
equipment
data center
neural network
machine room
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202011332596.7A
Other languages
Chinese (zh)
Other versions
CN112464311B (en
Inventor
董文雷
张军
李仲华
马云飞
宋扬
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhongtian Zhongda Smart City Technology Co ltd
Original Assignee
Zhongtian Zhongda Smart City Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zhongtian Zhongda Smart City Technology Co ltd filed Critical Zhongtian Zhongda Smart City Technology Co ltd
Priority to CN202011332596.7A priority Critical patent/CN112464311B/en
Publication of CN112464311A publication Critical patent/CN112464311A/en
Application granted granted Critical
Publication of CN112464311B publication Critical patent/CN112464311B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/047—Probabilistic or stochastic networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G06Q10/043—Optimisation of two dimensional placement, e.g. cutting of clothes or wood

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Software Systems (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Computational Linguistics (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Business, Economics & Management (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Human Resources & Organizations (AREA)
  • Geometry (AREA)
  • Computer Hardware Design (AREA)
  • Probability & Statistics with Applications (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Physiology (AREA)
  • Development Economics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Computational Mathematics (AREA)
  • Genetics & Genomics (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Analysis (AREA)
  • Game Theory and Decision Science (AREA)
  • Medical Informatics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)

Abstract

The invention discloses an equipment arrangement method of a smart city data center, which comprises the following steps: step S1, abstracting the plane model of the machine room equipment into an abstract polygon in an overlook state, and recording the abstract polygon into a template library; step S2, obtaining position information of boundaries of each room of the data center; step S3, identifying the room name according to the characteristics of the space by using a depth vision method, specifically including: establishing a training set and a testing set, wherein the training set comprises training pictures and labels corresponding to the training pictures; constructing a deep neural network, wherein the deep neural network comprises an input layer, three hidden layers and an output layer, and the layers are fully connected; and step S4, performing machine room equipment arrangement by using multiple genetic algorithms. The invention quickly generates the layout scheme of the equipment room in batch, which meets the function judgment, novelty judgment and priori knowledge of the layout rule.

Description

Equipment arrangement method of smart city data center
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to an equipment arrangement method of a smart city data center.
Background
With the development of computer and network technologies, IT equipment such as servers and network communication equipment is developing towards miniaturization, networking and rack-based, and various equipment such as a plurality of data storage cabinets, heat dissipation devices and the like need to be placed in a data center machine room of a smart city, which is the core of comprehensive wiring and information-based network equipment and is also a data aggregation center of an information network system.
Data storage enclosures have long been regarded as low value, add-on products for servers in IT applications and are not considered as a significant concern. However, a low-value data storage cabinet is the most direct physical protection of expensive IT equipment, and attaches importance to the method that the IT equipment ignores the IT microenvironment where the IT equipment is located, such as heat management, cable management, cabinet power distribution, compatibility and other advanced performances, improper cabinet arrangement can cause high failure rate of the data storage cabinet of the data center, performance reduction, resource waste and the like.
Disclosure of Invention
In order to solve the above problems, the present invention provides an equipment arrangement method for a smart city data center, comprising the following steps: step S1, abstracting the plane model of the machine room equipment into an abstract polygon in an overlook state, and recording the abstract polygon into a template library; step S2, obtaining position information of boundaries of each room of the data center; step S3, identifying the room name according to the characteristics of the space by using a depth vision method, specifically including: establishing a training set and a testing set, wherein the training set comprises training pictures and labels corresponding to the training pictures; constructing a deep neural network, wherein the deep neural network comprises an input layer, three hidden layers and an output layer, all layers are connected, and j-th neurons of an r-th layer of the deep neural network are represented as follows:
Figure BDA0002796245000000021
wherein, the r-1 layer contains m neurons,
Figure BDA0002796245000000022
represents the kth neuron of the r-1 layer
Figure BDA0002796245000000023
To
Figure BDA0002796245000000024
The weight of (a) is determined,
Figure BDA0002796245000000025
is shown to
Figure BDA0002796245000000026
The offset of (a) is an activation function, w is a linear relation coefficient, b is a bias, the output layer adopts a softmax classifier, and the index of the neuron with the maximum probability value is output as a prediction result of the neural network on the room picture; and step S4, performing machine room equipment arrangement by using multiple genetic algorithms.
According to the invention, multiple genetic algorithms are adopted, a plurality of machine room equipment distribution schemes are generated by utilizing the genetic algorithms for the whole data center, the genetic algorithms are utilized to calculate each machine room equipment distribution scheme according to each room, so that the layout scheme of each room is obtained, and the machine room equipment layout scheme which accords with layout rule function judgment, novelty judgment and priori knowledge can be rapidly generated in batches.
Drawings
The above features and technical advantages of the present invention will become more apparent and readily appreciated from the following description of the embodiments thereof taken in conjunction with the accompanying drawings.
FIG. 1 is a schematic diagram illustrating the steps of a method for arranging devices in a smart city data center according to an embodiment of the present invention;
fig. 2 is a schematic diagram illustrating an arrangement of racks in a data center according to an embodiment of the invention.
Detailed Description
Embodiments of the present invention will be described below with reference to the accompanying drawings. Those of ordinary skill in the art will recognize that the described embodiments can be modified in various different ways, or combinations thereof, without departing from the spirit and scope of the present invention. Accordingly, the drawings and description are illustrative in nature and not intended to limit the scope of the claims. Furthermore, in the present description, the drawings are not to scale and like reference numerals refer to like parts.
Fig. 1 is a flowchart illustrating an apparatus arrangement method of a smart city data center according to the present invention. An embodiment of the present invention is described below with reference to fig. 1.
The equipment arrangement method of the smart city data center comprises the following steps:
and step S1, abstracting the plane models (under the overlooking condition) of some common machine room equipment into corresponding abstract polygons, and recording the abstract polygons into a template library. For example, the planar model of a square device from the top is a square and the planar model of a rectangular device is a rectangle. Of course this embodiment does not exclude other abstract polygon forms which may also be more complex.
And step S2, obtaining the boundary information of the data center room. The data center may be an entire room or may be a plurality of connected rooms. The machine room plane graph can be converted into a gray scale graph by using a binarization method, and then the boundary of each room is extracted by using a flooding method.
The binarization is to convert the machine room plane graph into a grey-scale graph by setting a grey-scale value threshold, comparing the pixel values of all points of the machine room plane graph with the grey-scale value threshold, setting the pixel values of the pixel points which are greater than or equal to the grey-scale value threshold to be 255, and setting the pixel values of the pixel points which are smaller than the grey-scale value threshold to be 0. The gray value threshold may be a median value selected from 0 to 255.
In which the flooding method acquires each connected domain (i.e. each room region) in the gray-scale map, and extracts these regions by labeling. Calculating the connected domain is to check the connectivity of each pixel with its neighboring pixels. The scanning can be performed from left to right, and then the line is changed downwards, and the scanning is continued from left to right, and when one pixel is scanned, the adjacent pixel values of the pixel positions, namely the upper, lower, left and right, are checked, or the adjacent pixel values of the upper, lower, left, right, upper left, upper right, lower left and lower right are checked.
The following specific steps are described by taking the above, below, left and right inspections as examples:
assuming that the pixel value of the current position is 255, two adjacent pixels to its left and top (which must be scanned before the current pixel) are examined. The two combinations of pixel values and labels have the following four cases:
1) if the pixel values at the left and top are both 0, then a new label is given to the pixel at the current position (indicating the start of a new connected component);
2) only one pixel value on the left side and the upper side is 255, and the mark of the pixel at the current position is the same as that of the pixel with the pixel value of 255;
3) the pixel values on the left and the top are 255 and the labels are the same, so that the labels of the pixels at the current position are the same as those of the pixels on the left and the top;
4) if the pixel values on the left side and the top side are 255 and the labels are different, the smaller label is assigned to the pixel at the current position, and then the tracing back is performed from right to left until the starting pixel of the area, and the 4 steps are respectively performed each time of tracing back.
The regions of each room can be extracted from the gray-scale map through the above 4 steps.
And step S3, automatically identifying the space name according to the characteristics of the space (such as floor, flue, direction, area and the like) by using a depth vision method, and writing a rule function to correct the space name by combining a plurality of priori knowledge. For example, the deep vision method employs a neural network model to extract features of a room, and then performs classification recognition, thereby determining naming. The method specifically comprises the following steps:
step S31, a training set and a test set are established, where the training set includes training pictures and their corresponding real labels, each training picture is a top view of some rooms, and features of each room include, for example, a position of a power interface, a window orientation, a room size, and the like. The test set contains test images, and the pictures in the test set are predicted by training the pictures in the training set.
And step S32, constructing a deep neural network, wherein the deep neural network comprises an input layer, three hidden layers and an output layer. The layers are all connected, and a linear relation exists between the output and the input of each layer, and j < th > neuron of r < th > layer of the deep neural network is represented as follows:
Figure BDA0002796245000000041
wherein layer r-1 comprises m neurons;
Figure BDA0002796245000000042
represents the kth neuron of the r-1 layer
Figure BDA0002796245000000043
To
Figure BDA0002796245000000044
The weight of (c);
Figure BDA0002796245000000045
is shown to
Figure BDA0002796245000000046
The offset of (a);
σ (-) is the activation function;
w is a linear relation coefficient;
b is bias;
and the output layer adopts a softmax classifier and outputs the index of the neuron with the maximum probability value as a prediction result of the neural network on the room picture.
Step S4, machine room equipment arrangement is carried out by using multiple genetic algorithms, wherein the multiple genetic algorithms comprise a genetic algorithm for obtaining machine room equipment distribution of the whole data center machine room and a genetic algorithm for machine room equipment arrangement of each room, and specifically, the method comprises the following sub-steps:
step S41, forming a first gene code for each equipment room (including different types, shapes and sizes), so as to establish a mapping relationship between the equipment room and the gene.
Step S42, a population is initialized at random, where the population includes a plurality of computer room equipment allocation schemes, each scheme is to allocate different computer room equipment to each room of the whole data center computer room (but at this time, computer room equipment layout has not been performed yet), each scheme corresponds to a first gene encoding string composed of first gene codes, and each first gene encoding string may be referred to as a gene individual in the population in the genetic algorithm.
In step S43, a fitness evaluation is performed on each genetic individual by using a first fitness function, which contains a priori knowledge of the entire set of rooms, for example, a data center generally includes a main room for storing cabinets, a basic working room (including offices, dressing rooms, etc.), a first type of auxiliary room (including maintenance rooms, instrument rooms, spare parts rooms, storage medium storage rooms, data rooms), a second type of auxiliary room (including low-voltage power distribution, UPS power supply rooms, storage battery rooms, precision air conditioning system rooms, gas fire extinguishing equipment rooms, etc.), and a third type of auxiliary room (including storage rooms, general rest rooms, hand washing rooms, etc.).
According to the matching of each device and each room, for example, a main computer room usually has a cabinet, an air conditioner, a UPS, a power distribution cabinet, an automatic fire alarm device, etc., a maintenance room usually has a corresponding maintenance device, and a rest room usually has a bed. And forming a judgment condition by the prior knowledge to form a first adaptive function, and eliminating the cabinet distribution scheme lower than the fitness threshold value through the grading of the first adaptive function. The first adaptive function can also be set according to own preference as a priori knowledge, so that the arrangement of the first adaptive function is more in line with own will.
And step S44, for the schemes which do not reach the score, carrying out combined intersection and mutation, wherein the combined intersection refers to exchanging one or more genes in one layout with genes in the corresponding position in another layout so as to obtain a new overall first gene coding string. The mutation is to perform mutation operation on the gene coding string according to mutation probability and randomly assigned certain gene or certain genes, so as to form a new whole gene coding string. Corresponding to the binary code string, a certain bit or a plurality of bits randomly assigned are converted into '1' and '0'. For example, after the binary code 01_11_01_10_10_11 is mutated, the binary code may become a new code string 00_11_01_01_10_11, and the new first gene code string is executed again in the iterative loop of S43 and S44, so as to gradually increase the score of the fitness function until the fitness threshold is reached;
step S45, a plurality of machine room equipment allocation plans meeting the fitness threshold are generated.
Then, at step S5, using a genetic algorithm again for each room using parallel calculations to evolve, corresponding to any of the equipment room allocation schemes, the layout of the equipment room is determined. Specifically corresponding to each room, the method comprises the following substeps:
step S51, designing a genetic code of the genetic algorithm, and using an abscissa x, an ordinate y, and a rotation angle α of the machine room device in a horizontal plane of the center of the abstract polygon of the machine room device in the room as a second genetic code of the machine room device, where the abscissa x, the ordinate y, and the rotation angle α may be formed by binary coding. And respectively forming second gene codes for the positions of all the equipment in the machine room, and arranging the second gene codes in a row according to the sequence, such as: x1_ y1_ α 1_ x2_ y2_ α 2_ _______________________ xN _ α N, forming a second gene encoding string of the overall machine room equipment layout. The initial values of the abscissa x, the ordinate y and the rotation angle α may be set manually, for example, the initial values of x and y are set such that the boundary between the abstract polygon and the room reaches a set distance, and the initial value of α is set to 0.
For each room, a plurality of second gene-encoded strings are randomly generated as an initial population, the second gene-encoded strings being represented as:
Figure BDA0002796245000000061
wherein, the vector kmThe m-th second gene encoding string representing the room,
Figure BDA0002796245000000062
and a second gene code of j machine room equipment of the m second gene code string in the population is represented, and N represents the total number of the machine room equipment in one room. For example, one second gene encoding string may be a cabinet combining 24U (1200mm x 600mm), 32U (1600mm x 600mm), 42U (2000mm x 600mm) … … arranged in sequence with coordinate and angle information.
Step S52, machine room equipment layout rule functions of each space are designed, the machine room equipment layout rule functions comprise function dependency, orientation constraint and basic standard constraint among all machine room equipment, and the more the machine room equipment layout rule functions conform to the standard constraint, the higher score is given. For example, the equipment arrangement of the machine room is adapted to the position of a power supply so as to facilitate the laying of strong and weak cables; for example, machine room equipment of the same brand tends to be arranged in the same area, which facilitates maintenance operation; for example, the equipment rooms are arranged back to back and face to face, so that cold air and hot air are effectively separated, and the clear width d of the passageway between two rows of equipment rooms is not less than 1.2 m.
For the rule function in terms of distance, a distance threshold may be set, for example, the width of an aisle between two rows of cabinets is calculated by using the position coordinates of the cabinet 100, and if the distance is lower than the set distance threshold, it is determined that the specification constraint is not met. Or the distance between the machine room equipment and the wall surface can be judged whether to accord with the standard constraint or not through a set distance threshold value. If the distance threshold is exceeded, the specification constraint is determined not to be met, and a layout scheme is rejected.
For orientation constraints, the cabinet should be back-to-back and face-to-face as examples. Firstly, whether the rotation angles alpha of the cabinets are consistent or not is verified, and if not, the layout scheme is directly rejected. If the two are consistent, whether the back-to-back and face-to-face characteristics between the cabinet and the cabinet are met is further judged.
Several bits can be added in the second gene code of the cabinet to represent the position coordinates of the front and the back of the cabinet, and the position coordinates of the front of two adjacent cabinets are used for judging whether the prior knowledge is met. For example, for a second gene encoding string x1_ y1_ α 1_ a _ x2_ y2_ B _ x3_ y3_ x4_ y4_ α 4_ a _ x5_ y5_ B _ x6_ y6, wherein the second gene encoding x1_ y1_ α 1_ a _ x2_ y2_ B _ x3_ y3 represents the first cabinet, x1_ y1_ α 1 represents the coordinate and rotation angle of the center of the first cabinet, a _ x2_ y2 represents the coordinate of the center of the front of the first cabinet, and B _ x3_ y3 represents the coordinate of the center of the back of the first cabinet.
x4_ y4_ α 4_ A _ x5_ y5_ B _ x6_ y6 represents the second cabinet, and likewise, x4_ y4_ α 4 represents the center coordinates and the rotation angle of the second cabinet. A _ x5_ y5 represents the coordinates of the center of the front of the second cabinet, and B _ x6_ y6 represents the coordinates of the center of the back of the second cabinet.
Whether the two second cabinets are opposite in back is judged through the following formula:
Z1＝Q1+Q2+P1
where P1 is the distance between the back of two cabinets,
z1 is the distance between the two cabinet fronts;
q1 is the width between the front and back of a cabinet;
q2 is the width between the front and back of the other cabinet.
Step S53, a novelty function is designed, which means that the more novel layout is given a higher score. The method for judging novelty is to calculate the novelty by comparing the similarity between the generated layout and the existing and existing layout, and the higher the similarity with the existing layout, the lower the novelty, the specific operations are as follows:
step S531, extracting the machine room equipment and the orientation of the machine room equipment in the existing indoor layout by identifying the existing indoor layout, and forming a comparison gene library containing multiple machine room equipment layouts V1, V2 … and Vn in the same second gene coding mode as the step S51, wherein V1, V2 … and Vn are second gene coding strings respectively;
step S532, when calculating the novelty of a layout, normalizing the second gene encoding strings V1, V2 …, Vn of the comparison gene library and the second gene encoding string B of the new layout to be compared, respectively;
step S533, calculating the azimuth difference vector δ as V1-B, representing the degree of difference between the second gene encoding string V1 and the second gene encoding string B by norm | | δ | |, and similarly calculating the degree of difference between the second gene encoding strings V2 through An and the second gene encoding string B newly laid out, to obtain the sum of the degrees of difference between B and V1, V2 …, Vn, and the higher the sum of the degrees of difference is, the higher the novelty is, the higher the score is given.
In step S54, the above series of machine room equipment layout rule functions and novelty functions are combined with corresponding weights into an adaptive function in a genetic algorithm.
Specifically, f (g) ═ E1*G1+E2*G2…Ei*Gi+En-1*Gn-1+En*Gn+E*F
Wherein f (G) is an adaptive function;
Giis a function of the ith machine room equipment layout rule, EiIs GiThe corresponding weight value can be set manually according to the preference;
n represents the number of the layout rule functions of the equipment in the machine room;
f is a novelty function; and E is the weight corresponding to F.
Therefore, for a computer room equipment layout, a second gene coding string is input into the adaptability function, and a score can be obtained. For each room, a score, i.e., an adaptive score, is obtained for the different layouts of the room.
In step S55, it is determined whether a fitness threshold is reached corresponding to each room, and if the fitness threshold is reached in each room, the evolution results of the rooms are combined to form an optimal machine room equipment layout, and step S6 is performed. If the fitness of the room does not reach the fitness threshold, the following step S56 is performed.
In step S56, for the room that does not reach the fitness threshold, performing combinatorial crossing and mutation, and performing an iterative loop of step S54 and step S55 on the new second gene encoding string again, thereby gradually increasing the score of the fitness function until the fitness threshold is reached;
in step S6, the generated first gene encoding string of the optimal layout of the computer room equipment of each evolution instance is decoded into an abstract polygon of the specific computer room equipment, the model library is queried, the second gene encoding string is decoded into the position parameters of the layout of the computer room equipment, the corresponding plane model is selected and drawn into the computer room plane graph, and a series of corresponding output files are generated.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (1)

1. A device arrangement method of a smart city data center is characterized by comprising the following steps:
step S1, abstracting the plane model of the machine room equipment into an abstract polygon in an overlook state, and recording the abstract polygon into a template library;
step S2, obtaining position information of boundaries of each room of the data center;
step S3, identifying the room name according to the characteristics of the space by using a depth vision method, specifically including:
establishing a training set and a testing set, wherein the training set comprises training pictures and labels corresponding to the training pictures;
constructing a deep neural network, wherein the deep neural network comprises an input layer, three hidden layers and an output layer, all layers are connected, and j-th neurons of an r-th layer of the deep neural network are represented as follows:
Figure FDA0002796244990000011
wherein, the r-1 layer contains m neurons,
Figure FDA0002796244990000012
represents the kth neuron of the r-1 layer
Figure FDA0002796244990000013
To
Figure FDA0002796244990000014
The weight of (a) is determined,
Figure FDA0002796244990000015
is shown to
Figure FDA0002796244990000016
The offset of (a) is an activation function, w is a linear relation coefficient, b is a bias, the output layer adopts a softmax classifier, and the index of the neuron with the maximum probability value is output as a prediction result of the neural network on the room picture;
and step S4, performing machine room equipment arrangement by using multiple genetic algorithms.
CN202011332596.7A 2020-11-24 2020-11-24 Equipment arrangement method for smart city data center Active CN112464311B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011332596.7A CN112464311B (en) 2020-11-24 2020-11-24 Equipment arrangement method for smart city data center

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011332596.7A CN112464311B (en) 2020-11-24 2020-11-24 Equipment arrangement method for smart city data center

Publications (2)

Publication Number Publication Date
CN112464311A true CN112464311A (en) 2021-03-09
CN112464311B CN112464311B (en) 2024-08-13

Family

ID=74798875

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011332596.7A Active CN112464311B (en) 2020-11-24 2020-11-24 Equipment arrangement method for smart city data center

Country Status (1)

Country Link
CN (1) CN112464311B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000285149A (en) * 1999-03-30 2000-10-13 Sekisui Chem Co Ltd House design system using genetic algorithm
JP2001084285A (en) * 1999-09-14 2001-03-30 Sekisui Chem Co Ltd House design system using genetic algorithm
WO2018170421A1 (en) * 2017-03-17 2018-09-20 Magic Leap, Inc. Room layout estimation methods and techniques
CN111598972A (en) * 2020-04-03 2020-08-28 深圳大学 House type graph generation method and device, computer equipment and storage medium

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000285149A (en) * 1999-03-30 2000-10-13 Sekisui Chem Co Ltd House design system using genetic algorithm
JP2001084285A (en) * 1999-09-14 2001-03-30 Sekisui Chem Co Ltd House design system using genetic algorithm
WO2018170421A1 (en) * 2017-03-17 2018-09-20 Magic Leap, Inc. Room layout estimation methods and techniques
CN111598972A (en) * 2020-04-03 2020-08-28 深圳大学 House type graph generation method and device, computer equipment and storage medium

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
HU RUIZHEN ETAL: "《Graph2Plan: Learning Floorplan Generation from Layout Graphs》", 《ACM TRANS. GRAPH》, vol. 39, no. 4, 27 April 2020 (2020-04-27), pages 1 - 14 *

Also Published As

Publication number Publication date
CN112464311B (en) 2024-08-13

Similar Documents

Publication Publication Date Title
CN112464309B (en) Equipment arrangement method for smart city data center
WO2021155650A1 (en) Image recognition model training method and apparatus, computer system, and storage medium
CN110781406B (en) Social network user multi-attribute inference method based on variational automatic encoder
CN109635056B (en) Power utilization address data processing method and device, computer equipment and storage medium
CN106060008B (en) A kind of network intrusions method for detecting abnormality
CN109829065B (en) Image retrieval method, device, equipment and computer readable storage medium
WO2016095068A1 (en) Pedestrian detection apparatus and method
CN114168795B (en) Building three-dimensional model mapping and storing method and device, electronic equipment and medium
CN113688797A (en) Abnormal behavior identification method and system based on skeleton extraction
CN109948609A (en) Intelligently reading localization method based on deep learning
CN112464312B (en) Equipment arrangement method for smart city data center
CN113836707A (en) Electric power system community detection method and device based on accelerated attribute network embedding algorithm
CN116884192A (en) Power production operation risk early warning method, system and equipment
CN114201161A (en) Module division method for graphical programming
CN110659997A (en) Data cluster identification method and device, computer system and readable storage medium
CN112464311B (en) Equipment arrangement method for smart city data center
CN112464310B (en) Equipment arrangement method for smart city data center
CN116522565B (en) BIM-based power engineering design power distribution network planning method and computer equipment
CN116665482B (en) Parking space recommending method and device based on intelligent parking
CN112419243A (en) Power distribution room equipment fault identification method based on infrared image analysis
CN116911332A (en) Classified storage method, device, equipment and medium based on cargo characteristics
CN115049836A (en) Image segmentation method, device, equipment and storage medium
CN109635057B (en) Electricity utilization business processing method and device, computer equipment and storage medium
CN111191717B (en) Black box confrontation sample generation algorithm based on hidden space clustering
CN114648762A (en) Semantic segmentation method and device, electronic equipment and computer-readable storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant