CN114117238A - Intelligent caregiver recommendation scheduling system - Google Patents

Intelligent caregiver recommendation scheduling system Download PDF

Info

Publication number
CN114117238A
CN114117238A CN202111499182.8A CN202111499182A CN114117238A CN 114117238 A CN114117238 A CN 114117238A CN 202111499182 A CN202111499182 A CN 202111499182A CN 114117238 A CN114117238 A CN 114117238A
Authority
CN
China
Prior art keywords
caregiver
neural network
scheduling
weight
time
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
CN202111499182.8A
Other languages
Chinese (zh)
Other versions
CN114117238B (en
Inventor
张军
袁翔
赵鸿飞
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.)
Fushoukang Shanghai Medical And Pension Service Co ltd
Original Assignee
Fushoukang Shanghai Medical And Pension Service 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 Fushoukang Shanghai Medical And Pension Service Co ltd filed Critical Fushoukang Shanghai Medical And Pension Service Co ltd
Priority to CN202111499182.8A priority Critical patent/CN114117238B/en
Priority to CN202311320590.1A priority patent/CN117393124A/en
Publication of CN114117238A publication Critical patent/CN114117238A/en
Application granted granted Critical
Publication of CN114117238B publication Critical patent/CN114117238B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
    • G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
    • 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
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06311—Scheduling, planning or task assignment for a person or group
    • G06Q10/063112—Skill-based matching of a person or a group to a task
    • 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
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06398—Performance of employee with respect to a job function
    • 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
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/26—Government or public services
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Abstract

The invention provides an intelligent caregiver recommendation scheduling system, which comprises: an information extraction module configured to extract caregiver features of the respective caregivers and caregivers features of the caregivers; the neural network construction module is configured to construct a neural network according to the characteristics of the caregivers and the characteristics of the cared persons, the historical data of the characteristics of the caregivers and the characteristics of the cared persons are trained, and recommendation scores of pairwise matching of the characteristics of the cared persons and the characteristics of the caregivers are obtained, and the recommendation scores are degrees to which the corresponding caregivers are to be allocated to the corresponding cared persons; and the scheduling module is configured to rank the scores of the qualified nursing staff and recommend the scores to the scheduling staff, and the scheduling staff preferably selects the score to be the highest.

Description

Intelligent caregiver recommendation scheduling system
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to an intelligent caregiver recommendation scheduling system.
Background
According to the research on the aged market and the development trend of the Chinese population of 2019-2025, according to the international traffic division standard, when the proportion of the population of 65 years old or more in a country or a region exceeds 7 percent, the aging is realized; the aging rate reaches 14 percent, which is the deep aging; if the content exceeds 20%, the people enter a super-aging society. As can be seen from the second figure, China has entered the aging stage at present. The population proportion of the aged 65 years old and older in China reaches 11.70 percent in 2020, and the people are about to enter deep aging. The proportion of the population of the aged 65 years old and older in 2040 years old is estimated to exceed 20 percent, and the aged people enter into the super-aging society.
The existing home care market is large in scale, wide in geographical range and dispersed in service, and can not realize uniform care and uniform management similar to care institutions; on the other hand, according to relevant data statistics, the aged care nursing staff in the Chinese aged care industry needs 1300 ten thousand of employees, but the actual employees are less than 30 ten thousand, so that the development of the aged care industry is more limited, and the aged care needs of a plurality of old people cannot be met.
Furthermore, Chinese nursing staff has huge population defects, and the period for cultivating the nursing staff is long. After a nursing staff passes through long-time training and the 'career qualification examination' of an aged nursing staff, the nursing staff can take the qualification certificate of the nursing staff and then perform the certification for post work. Not only that, but also because of the manual limitations leading to unreasonable work schedules, a large number of caregivers are not reasonably allocated, which further exacerbates the shortage of caregivers.
Disclosure of Invention
The invention aims to provide an intelligent caregiver recommendation scheduling system to solve the problem that the influence caused by shortage of caregivers is more serious due to unreasonable work allocation of the existing caregivers.
In order to solve the above technical problem, the present invention provides an intelligent caregiver recommendation scheduling system, which includes:
an information extraction module configured to extract caregiver features of a caregiver and caregiver-addressed features of a caregiver;
a neural network building module configured to perform the following actions:
constructing a neural network according to the characteristics of the caregiver and the characteristics of the cared person; and
training historical data of caregiver characteristics and caregiver characteristics to derive recommendation scores for pairwise caregiver characteristics and caregiver characteristics, wherein the recommendation scores characterize the degree to which a corresponding caregiver should be assigned to a corresponding caregiver;
a scheduling module configured to perform the following actions:
ranking the caregiver according to the recommendation score; and
and scheduling the nursing staff according to the ranking to nurse the cared person.
Optionally, in the caregiver intelligent recommendation scheduling system,
the information extraction module is further configured to extract existing and future work features of the caregiver; and/or
The neural network building module comprises a mathematical model building module, and the mathematical model building module is configured to obtain the most appropriate scheduling scheme function through data analysis and machine learning according to the existing working characteristics and the future working characteristics; and/or
The scheduling module is further configured to perform the following actions:
scheduling future work for the idle personnel according to the scheduling scheme function; and
and obtaining qualified nursing staff according to the schedule of the future work.
Optionally, in the caregiver intelligent recommendation scheduling system, the caregiver characteristics include one or more of the following: caregiver gender, caregiver age, caregiver category, caregiver work experience, company work experience, caregiver staff star, caregiver staff rating, number of complaints the caregiver has, total caregiver order, caregiver maximum scholastic calendar, caregiver political aspect, caregiver job type, caregiver marital status, caregiver height, caregiver weight, caregiver current order, caregiver last order time, caregiver next start time, caregiver last distance, caregiver next distance.
Optionally, in the caregiver intelligent recommendation scheduling system, the caregiver-related characteristics include one or more of the following: the sex of the cared, the age of the cared, the length of time the cared has signed up, the health grade of the cared, the average value of the grades of the cared, the number of complaints of the cared, the number of times the cared is served by the current cared and the number of times the cared is served by the current cared in 30 days.
Optionally, in the caregiver intelligent recommendation scheduling system, the system further includes:
the preprocessing module is configured to input the preprocessed caregiver characteristics and/or the preprocessed caregiver characteristics into the neural network construction module; the pretreatment comprises the following steps:
the sex of the caregiver and the sex of the caretaker are G e { M, F }, and specific values of sex input are calculated:
Figure BDA0003400790260000031
Ni＝1-M
where M represents all orders, n represents female employee sheets, SiA score representing the ith work order;
preprocessing the gender input by calculating the specific value of the gender input, and distributing a caregiver with proper gender to a cared person;
inputting the characteristics of the caregiver class according to the requirements of the carereceiver:
Figure BDA0003400790260000032
where r is the requirement of the care recipient, piIs a caregiver class;
the caregiver station types include full-time, part-time, practice and retirement, the caregiver station type assigned preferentially is full-time, and the average single formula for each caregiver station type is:
Figure BDA0003400790260000033
wherein μiAverage unit amount, count (S) representing the ith caregiver typei) Represents the number of employees of the ith type; order (S)ij) Representing the order quantity of the jth caregiver of the ith type caregiver; after the average single amount of the caregivers of each post type is calculated, the average single amount is substituted into a softmax function, the softmax function is standardized, and meanwhile, the weight value of the dominant post is amplified:
Figure BDA0003400790260000041
wherein ,SiRepresenting the characteristic input of the i-th type of caregiver.
Optionally, in the caregiver intelligent recommendation scheduling system, the preprocessing further includes:
calculating the time difference between the next order according to the previous single-ending time and the next single-starting time:
Figure BDA0003400790260000042
wherein TeIndicates the end time of the previous order, TsIndicates the start time, T, of the next orderiRepresenting a threshold time difference, if the time difference between the upper bill and the lower bill exceeds the threshold time difference, considering that a new bill can be inserted between the upper bill and the lower bill, and defining the threshold time difference according to the service condition;
Wtrepresenting the expected time of flight, PLNA feature input representing a time from a previous single bundle or a feature input representing a next single start time;
calculating the predicted journey time of the caregiver according to the predicted journey time from the previous order to the current order and the predicted journey time from the current order to the next order:
Figure BDA0003400790260000043
wherein (xi,yi) and (xj,yj) Respectively, longitude and latitude coordinate values of two places, v is the estimated speed of a vehicle used by a caregiver, and ρ is time redundancy which is preset according to the actual condition of the caregiver; the normalization of the time required for the journey is normalized using the modified mean variance:
Figure BDA0003400790260000044
where X is the time required for the trip, μ is the average of the caregiver's trip times, and σ is the variance of the caregiver's trip times.
Optionally, in the caregiver intelligent recommendation scheduling system, the neural network building module is further configured to:
constructing a BP (back propagation) neural network according to the preprocessed input characteristics of the caregiver and the characteristics of the caregivers, wherein the number of hidden layers of the BP neural network is one layer, an output layer of the BP neural network is a single neuron, and both the hidden layers and the output layer of the BP neural network adopt a Sigmoid function as an activation function;
after a BP neural network is constructed, a weight is given to a first layer neural network through an initialization weight module;
the initialization weight module adopts an expert system to calculate the weight of the first layer neural network, and the expert system adopts a large-scale multi-expert cooperation system, multi-knowledge representation, an integrated knowledge base, a self-organization problem solving mechanism, multi-subject cooperation problem solving and parallel reasoning, an expert system tool and environment and an artificial neural network knowledge acquisition and learning mechanism so as to have the weight of the first layer neural network with multiple knowledge bases and multiple main bodies.
Optionally, in the caregiver intelligent recommendation scheduling system, the weights of the first layer neural network include:
the highest scholastic calendar of the caregiver, the political face of the caregiver, the marital status of the caregiver, the height of the caregiver, the weight of the caregiver, the gender of the caregiver and the gender of the cared person are respectively 0.1;
the weight value of the health grade of the nursed person is 0.2;
the type of the caregiver, the type of the caregiver's posts, the working experience of the caregiver, and the weight of the signed time length of the cared person are 0.3;
the weight of the working experience of the company is 0.4;
the star rating of the nursing staff, the number of complaints of the nursing staff, the total single amount of the nursing staff, the age of the nursed person, the average value of the ratings of the nursed person and the number of complaints of the nursed person are 0.5;
the weight of the daily sheet amount of the nursing staff is 0.6;
the last single-ending time of the caregiver, the next starting time of the caregiver and the number of times of the current caregiver service of the cared person for the last 30 days are weighted to be 0.8;
the weight of the previous single distance of the caregiver, the next single distance of the caregiver, and the number of times the cared person is served by the current caregiver is 0.7.
Optionally, in the caregiver intelligent recommendation scheduling system, historical data is collected, the historical scheduling data is input into a neural network for learning, a final model is obtained for subsequent scheduling, after a worker schedules, new data is continuously input into an original model for training, and the model is continuously optimized;
when the station leader needs to arrange nurses for the nursed persons, the nursing persons meeting the conditions are inquired, the data of the nursing persons and the nursed persons meeting the conditions form a vector to be used as the input of a model, the obtained output is sequenced and recommended to the station leader, and the station leader selects the first nursing person for scheduling;
the input characteristics of the cared person are preprocessed, wherein the weights calculated by the initialization module are obtained by carrying out pre-judgment by an expert system according to experience, the iteration times of training are reduced, and the model training speed is accelerated.
After characteristics of a caregiver and a nursed person are collected, a BP neural network is constructed, the number of hidden layers is one, and an output layer is a single neuron; the hidden layer and the output layer both adopt Sigmoid functions as activation functions.
Optionally, in the caregiver intelligent recommendation scheduling system, the training process of the model includes:
the feature input vector is:
x＝(x0，x1，x2，···，xp，xp+1，·，xn)T
wherein x1To xpCharacteristic of the caregiver, xp+1To x0Is a characteristic of the person to be cared for, x0Is a bias node; the weights of the first layer neural network are:
Figure BDA0003400790260000061
wherein wijRepresenting the weight from the jth node of the input layer to the ith node of the hidden layer, the output of the hidden layer is
Z＝Wx
Substituting it into Sigmoid function to get:
Figure BDA0003400790260000062
after hidden layer input is obtained, forward propagation is continued, the weight of the second layer neural network is in a random initialization mode, and the weight is expressed as:
V＝(v0，v1，v2，…，vn)
wherein viRepresenting the ith weight from the hidden layer node to the output layer node; the final output is then:
Figure BDA0003400790260000071
after obtaining the predicted value ^ y, defining a loss function by using MSE, wherein the loss function is as follows:
Figure BDA0003400790260000072
wherein m represents the total number of samples; the gradient descent is carried out on the neural network through back propagation, and the weight of the first layer of neural network is adjusted to be as follows:
Figure BDA0003400790260000073
where eta is learning rate, partial differential
Figure BDA0003400790260000074
Comprises the following steps:
Figure BDA0003400790260000075
and adjusting the weight of the first layer of neural network, and obtaining the weight according to a chain rule:
Figure BDA0003400790260000076
wherein ,
Figure BDA0003400790260000077
comprises the following steps:
Figure BDA0003400790260000078
after updating, completing a first round of circulation; the above circulation is repeatedly carried out until the iteration change of two times tends to 0, the iteration is stopped, and the model training is finished; for the next shift, after the staff schedules, new data is continuously put into the original model for training, and the model is continuously optimized;
after the training of the model is completed, when the shift is scheduled, the characteristics of the caregiver and the characteristics of the cared person are preprocessed, then the trained model is imported, and the caregiver score is obtained:
Scorei＝f(Xi，Xj)
wherein XiFeature vector, X, representing a caregiverjRepresenting the feature vector of the caregiver, fScore for a trained model functioniA score for i caregiver; and finally, selecting the caregiver with the highest score from the selectable caregivers.
The invention also provides a caregiver intelligent recommendation scheduling method, which comprises the following steps:
the information extraction module extracts the caregiver characteristics of the caregiver and the caretaker characteristics of the caretaker;
the neural network building block performs the following actions:
constructing a neural network according to the characteristics of the caregiver and the characteristics of the cared person; and
training historical data of caregiver characteristics and caregiver characteristics to derive recommendation scores for pairwise caregiver characteristics and caregiver characteristics, wherein the recommendation scores characterize the degree to which a corresponding caregiver should be assigned to a corresponding caregiver;
a scheduling module configured to perform the following actions:
ranking the caregiver according to the recommendation score; and
and scheduling the nursing staff according to the ranking to nurse the cared person.
The inventor of the invention finds that the difference between two service places is too far, so that most of time of a caregiver is spent on the journey, and the difference between two service times is too far, so that a large amount of time is spent by the caregiver when the caregiver waits for the service time at the gate of the old man, so that the work allocation of the existing caregiver is unreasonable, and the influence caused by shortage of the caregiver is more serious.
Based on the above insights, the invention provides an intelligent caregiver recommendation scheduling system, which establishes a mathematical model by extracting characteristics of a caregiver, characteristics of the old (characteristics of a nursed person), existing working characteristics and future working characteristics, through data analysis and machine learning, the most suitable scheduling scheme function is learned, and finally, the idle personnel are scheduled for future work according to the model, in addition, a neural network is constructed according to the characteristics of the nursing staff and the characteristics of the old people, historical data is put into training, a score between 0 and 1 is finally obtained, the score represents the degree of whether the caregiver should be assigned to the elderly, the qualified caregiver is ranked for score, recommended to the shift clerk, the shift clerk preferably selects the score to be the highest, the optimal solution of the nursing staff work allocation can be realized, and the defects caused by the short term of the nursing staff are greatly alleviated.
Drawings
FIG. 1 is a diagram of a neural network of a caregiver intelligent recommendation scheduling system in an embodiment of the invention.
Detailed Description
The invention is further elucidated with reference to the drawings in conjunction with the detailed description.
It should be noted that the components in the figures may be exaggerated and not necessarily to scale for illustrative purposes. In the figures, identical or functionally identical components are provided with the same reference symbols.
In the present invention, "disposed on …", "disposed over …" and "disposed over …" do not exclude the presence of an intermediate therebetween, unless otherwise specified. Further, "disposed on or above …" merely indicates the relative positional relationship between two components, and may also be converted to "disposed below or below …" and vice versa in certain cases, such as after reversing the product direction.
In the present invention, the embodiments are only intended to illustrate the aspects of the present invention, and should not be construed as limiting.
In the present invention, the terms "a" and "an" do not exclude the presence of a plurality of elements, unless otherwise specified.
It is further noted herein that in embodiments of the present invention, only a portion of the components or assemblies may be shown for clarity and simplicity, but those of ordinary skill in the art will appreciate that, given the teachings of the present invention, required components or assemblies may be added as needed in a particular scenario. Furthermore, features from different embodiments of the invention may be combined with each other, unless otherwise indicated. For example, a feature of the second embodiment may be substituted for a corresponding or functionally equivalent or similar feature of the first embodiment, and the resulting embodiments are likewise within the scope of the disclosure or recitation of the present application.
It is also noted herein that, within the scope of the present invention, the terms "same", "equal", and the like do not mean that the two values are absolutely equal, but allow some reasonable error, that is, the terms also encompass "substantially the same", "substantially equal". By analogy, in the present invention, the terms "perpendicular", "parallel" and the like in the directions of the tables also cover the meanings of "substantially perpendicular", "substantially parallel".
The numbering of the steps of the methods of the present invention does not limit the order of execution of the steps of the methods. Unless specifically stated, the method steps may be performed in a different order.
The caregiver intelligent recommendation scheduling system according to the present invention will be described in detail with reference to the accompanying drawings and specific embodiments. The advantages and features of the present invention will become more apparent from the following description. It is to be noted that the drawings are in a very simplified form and are not to precise scale, which is merely for the purpose of facilitating and distinctly claiming the embodiments of the present invention.
The invention aims to provide an intelligent caregiver recommendation scheduling system to solve the problem that the influence caused by shortage of caregivers is more serious due to unreasonable work allocation of the existing caregivers.
The existing artificial intelligence algorithm comprises an automatic dispatching algorithm of the taxi taking industry and an automatic dispatching algorithm of the takeaway industry. The taxi taking industry currently has a very mature order dispatching algorithm. The biggest principle of the dispatching algorithm in the taxi taking industry is the 'near principle', namely, the order is distributed to the drivers closest to the current user. Besides, the characteristics of the driver, such as the score, the single amount, the vehicle type and the like, and the characteristics of the user, such as the complaint amount, the destination and the like, are comprehensively compared, and finally, a suitable driver is selected for dispatching the order. The takeaway order dispatching algorithm automatically dispatches orders by constructing a machine learning model. The machine learning model automatically calculates a "meal time estimate", "delivery time estimate", "future order estimate", "route elapsed time estimate". The colleagues carry out 'path planning' and 'automatic order dispatching' according to the overall optimization field knowledge.
However, the inventor has found that the artificial intelligence algorithm in the prior art is not suitable for the caregiver to get on the door to dispatch, for example, compared with the automatic dispatch of the taxi, the dispatch in the taxi dispatching field is real-time dispatch, that is, when the user takes an order, the driver is arranged for the user according to the user situation. But the attendance service list of the nursing staff is sent in advance, the client and the enterprise sign a long-term contract, then the nursing staff is arranged to carry out the attendance service at a fixed certain time every week, and the characteristics required to be considered by driving are not suitable for the home-based care industry, can not be directly applied and can not be converted. In addition, the taxi taking and dispatching need not consider whether the driver skill and the user requirement are matched or not. However, in the home-based care industry, it is necessary to consider whether the caregiver has the skills required by the customer, such as massage and tuina. The last taxi taking is to start from the current position, carry out the next order, and the destination of the next station is uncertain. But the caregiver is starting from the last customer's home and the destination of the next order is determined.
Further, the inventor finds that the takeout worksheet requires considering the current quantity of the worksheets held by the takeout worksheets through comparison between the attendance worksheets and the takeout automatic worksheets of the nursing staff, and one takeout worksheet can hold multiple worksheets at the same time. But the nursing staff can only serve one sheet at a time, and when the single sheet is finished, the next sheet can be served, the takeout has high requirements on the path, but the nursing staff does not need to excessively analyze the service path of the nursing staff. After the take-out arrives at the customer's home, the service ends. But after the caregiver arrives at the customer's home, a certain service time is generated, and when the service time is over, the next order can be continued. In summary, the existing order dispatching algorithm cannot solve the scheduling problem of the caregiver well, and a scheduling method needs to be specified for the caregiver with a pertinence according to the working characteristics of the caregiver. The invention provides an intelligent caregiver recommendation and scheduling system, belongs to the technical field of home care and mainly aims at reasonably arranging caregiver work.
The invention discloses a method for intelligently scheduling future work of a caregiver according to the existing work arrangement of the caregiver. The scheduling method is only suitable for the home-based care service industry, and the scheduling personnel can perform the most appropriate scheduling operation on the self service personnel according to the rules provided by the scheduling method, so that the working time of the service personnel is reasonably distributed, the time fragment of the service personnel is reduced, and the profit of an enterprise is maximized. The scheme is characterized in that a mathematical model is established by extracting characteristics of a caregiver, existing working characteristics and future working characteristics, and the most appropriate scheduling scheme function is learned through data analysis and machine learning. And finally, scheduling future work for the idle personnel according to the model. The invention will be described in detail, including but not limited to: the invention background, the technical scheme, the beneficial effects, the applicable scene, the use description and the model explanation. By reading the text, the use mode of the invention can be mastered and applied to actual scenes.
In order to achieve the above object, the present invention provides an intelligent caregiver recommendation scheduling system, comprising: an information extraction module configured to extract caregiver features of the respective caregivers and caregivers features of the caregivers; the neural network construction module is configured to construct a neural network according to the characteristics of the caregivers and the characteristics of the cared persons, the historical data of the characteristics of the caregivers and the characteristics of the cared persons are trained, and recommendation scores of pairwise matching of the characteristics of the cared persons and the characteristics of the caregivers are obtained, and the recommendation scores are degrees to which the corresponding caregivers are to be allocated to the corresponding cared persons; and the scheduling module is configured to rank the scores of the qualified nursing staff and recommend the scores to the scheduling staff, and the scheduling staff preferably selects the score to be the highest.
The embodiment of the invention provides an intelligent caregiver recommendation scheduling system, which comprises: an information extraction module configured to extract caregiver features of the respective caregivers and caregivers features of the caregivers; the neural network construction module is configured to construct a neural network according to the characteristics of the caregivers and the characteristics of the cared persons, the historical data of the characteristics of the caregivers and the characteristics of the cared persons are trained, and recommendation scores of pairwise matching of the characteristics of the cared persons and the characteristics of the caregivers are obtained, and the recommendation scores are degrees to which the corresponding caregivers are to be allocated to the corresponding cared persons; and the scheduling module is configured to rank the scores of the qualified nursing staff and recommend the scores to the scheduling staff, and the scheduling staff preferably selects the score to be the highest.
In one embodiment of the invention, in the caregiver intelligent recommendation scheduling system, the information extraction module is further configured to extract existing work features and future work features of each caregiver; the neural network building module comprises a mathematical model building module, and the mathematical model building module is configured to obtain the most appropriate scheduling scheme function through data analysis and machine learning according to the existing working characteristics and the future working characteristics; a scheduling module further configured to perform the following actions: scheduling future work for the idle personnel according to the scheduling scheme function; and obtaining qualified nursing staff according to the schedule of the future work.
The invention constructs a neural network by extracting characteristics of a caregiver and characteristics of the old, puts historical data into training, and finally obtains a score between 0 and 1, wherein the score represents the degree of whether the caregiver should be assigned to the old. And (4) ranking the scores of the qualified nursing staff, and recommending the nursing staff to the shift scheduling staff, wherein the shift scheduling staff can select the score with the highest priority in principle.
In one embodiment of the invention, in the intelligent caregiver recommendation scheduling system, the caregiver characteristics include caregiver gender, caregiver age, caregiver type, caregiver work experience, company work experience, caregiver staff star rating, caregiver staff rating, number of complaints made by a caregiver, total caregiver sheet amount, caregiver maximum scholarness, caregiver political aspect, caregiver position type, caregiver marital status, caregiver height, caregiver weight, caregiver current sheet amount, caregiver last unionbundle time, caregiver next sheet start time, caregiver last sheet distance, and caregiver next sheet distance.
In one embodiment of the invention, in the caregiver intelligent recommendation scheduling system, the characteristics of the carereceiver comprise sex of the carereceiver, age of the carereceiver, length of time of sign contract of the carereceiver, health grade of the carereceiver, average value of grades of the carereceiver, complaints of the carereceiver, service times of the carereceiver by the current carereceiver and service times of the carereceiver by the current carereceiver in the last 30 days. In the elderly care industry, the characteristics to be considered are different from the traditional industry.
In an embodiment of the present invention, in the caregiver intelligent recommendation scheduling system, further includes: the preprocessing module is configured to input the preprocessed caregiver characteristics and/or the preprocessed caregiver characteristics into the neural network construction module; the pretreatment comprises the following steps: the sex of the caregiver and the sex of the caretaker are G e { M, F }, and specific values of sex input are calculated:
Figure BDA0003400790260000131
Ni＝1-M
where M represents all orders, n represents female employee sheets, SiA score representing the ith work order; preprocessing the gender input by calculating the specific value of the gender input, and distributing a caregiver with proper gender to a cared person; the staff post is a special characteristic, and the degree of difficulty is different in different posts, and the different demands of old man also can influence the selection of post. The input of the characteristics needs to be combined with the requirements of the old, and the characteristics of the types of the caregivers are input according to the requirements of the cared persons:
Figure BDA0003400790260000132
where r is the requirement of the care recipient, piIs a caregiver class; the caregiver station types include full-time, part-time, practice and retirement, wherein the full-time is the largest and is the caregiver type that should be preferentially assigned, the preferentially assigned caregiver station type is full-time, and the average sheet formula of each caregiver station type is:
Figure BDA0003400790260000133
wherein μiAverage unit amount, count (S) representing the ith caregiver typei) Represents the number of employees of the ith type; order (S)ij) Representing the order quantity of the jth caregiver of the ith type caregiver; after the average single amount of the caregivers of each post type is calculated, the average single amount is substituted into a softmax function, the softmax function is standardized, and meanwhile, the weight value of the dominant post is amplified:
Figure BDA0003400790260000134
wherein ,SiRepresenting the characteristic input of the i-th type of caregiver.
In one embodiment of the present invention, in the caregiver intelligent recommendation scheduling system, the preprocessing further includes: the last (next) order end (start) time. The list to be arranged for the caregiver needs to take into account the time of the list and the caregiver's previous and next lists, and it is not reasonable to simply bring its value directly into negative correlation, requiring special handling. Calculating the time difference between the next order according to the previous single-ending time and the next single-starting time:
Figure BDA0003400790260000141
wherein TeIndicates the end time of the previous order, TsIndicates the start time, T, of the next orderiRepresenting a threshold time difference, if the time difference between the upper bill and the lower bill exceeds the threshold time difference, considering that a new bill can be inserted between the upper bill and the lower bill, and defining the threshold time difference according to the service condition; t isiIndicating that the time interval is exceeded, it is assumed that a new sheet can be inserted between the sheets, the value being a constant defined according to the traffic situation.
WtRepresenting the expected time of flight, PLNFeature input or next sheet representing time to last single bundle from distanceInputting the characteristics of the starting time;
calculating the predicted journey time of the caregiver according to the predicted journey time from the previous order to the current order and the predicted journey time from the current order to the next order:
Figure BDA0003400790260000142
wherein (xi,yi) and (xj,yj) The longitude and latitude coordinate values of two places are respectively, v is the estimated speed of a vehicle used by a caregiver, ρ is time redundancy, a certain time fault tolerance is added for the caregiver, and the value needs to be specified by a service worker according to the actual condition of the caregiver, so the time redundancy is preset according to the actual condition of the caregiver; the normalization of the time required for the journey is normalized using the modified mean variance:
Figure BDA0003400790260000143
where X is the time required for the trip, μ is the average of the caregiver's trip times, and σ is the variance of the caregiver's trip times. The neural network input features are finally assembled as shown in table 1:
table 1: caregiver profile
Figure BDA0003400790260000151
The input features of the nursing staff are processed according to the table mode, wherein the initialization weight is pre-judged by the service staff according to experience (not strict, can be randomly modified according to subjective will), so that the iteration times of training can be reduced, and the model training speed is accelerated. The elderly are characterized as shown in table 2:
table 2: caregiver profile
Figure BDA0003400790260000161
In one embodiment of the present invention, in the caregiver intelligent recommendation scheduling system, the neural network building module is further configured to: constructing a BP (back propagation) neural network according to the preprocessed input characteristics of the caregiver and the characteristics of the caregivers, wherein the number of hidden layers of the BP neural network is one layer, an output layer of the BP neural network is a single neuron, and both the hidden layers and the output layer of the BP neural network adopt a Sigmoid function as an activation function; after a BP neural network is constructed, a weight is given to a first layer neural network through an initialization weight module; the initialization weight module adopts an expert system to calculate the weight of the first layer neural network, and the expert system adopts a large-scale multi-expert cooperation system, multi-knowledge representation, an integrated knowledge base, a self-organization problem solving mechanism, multi-subject cooperation problem solving and parallel reasoning, an expert system tool and environment and an artificial neural network knowledge acquisition and learning mechanism so as to have the weight of the first layer neural network with multiple knowledge bases and multiple main bodies.
In an embodiment of the present invention, in the caregiver intelligent recommendation scheduling system, the weights of the first layer neural network include: the highest scholastic calendar of the caregiver, the political face of the caregiver, the marital status of the caregiver, the height of the caregiver, the weight of the caregiver, the gender of the caregiver and the gender of the cared person are respectively 0.1; the weight value of the health grade of the nursed person is 0.2; the type of the caregiver, the type of the caregiver's posts, the working experience of the caregiver, and the weight of the signed time length of the cared person are 0.3; the weight of the working experience of the company is 0.4; the star rating of the nursing staff, the number of complaints of the nursing staff, the total single amount of the nursing staff, the age of the nursed person, the average value of the ratings of the nursed person and the number of complaints of the nursed person are 0.5; the weight of the daily sheet amount of the nursing staff is 0.6; the last single-ending time of the caregiver, the next starting time of the caregiver and the number of times of the current caregiver service of the cared person for the last 30 days are weighted to be 0.8; the weight of the previous single distance of the caregiver, the next single distance of the caregiver, and the number of times the cared person is served by the current caregiver is 0.7.
In one embodiment of the invention, in the intelligent caregiver recommendation scheduling system, historical data is collected, the historical scheduling data is put into a neural network for learning, a final model is obtained for subsequent scheduling, and after a worker schedules, new data is continuously put into an original model for training, so that the model is continuously optimized; when the station leader needs to arrange nurses for the nursed persons, the nursing persons meeting the conditions are inquired, the data of the nursing persons and the nursed persons meeting the conditions form a vector to be used as the input of a model, the obtained output is sequenced and recommended to the station leader, and the station leader selects the first nursing person for scheduling; the input characteristics of the cared person are preprocessed, wherein the weights calculated by the initialization module are obtained by carrying out pre-judgment by an expert system according to experience, the iteration times of training are reduced, and the model training speed is accelerated. After characteristics of a caregiver and a nursed person are collected, a BP neural network is constructed, the number of hidden layers is one, and an output layer is a single neuron; the hidden layer and the output layer both adopt Sigmoid functions as activation functions. The neural network is shown in fig. 1.
In one embodiment of the present invention, in the caregiver intelligent recommendation scheduling system, the training process of the model includes:
the feature input vector is:
x＝(x0，x1，x2，···xp，xp+l，·，xn)T
wherein x1To xpCharacteristic of the caregiver, xp+1To xnIs a characteristic of the person to be cared for, x0Is a bias node; the weights of the first layer neural network are:
Figure BDA0003400790260000171
wherein wijRepresenting the weight from the jth node of the input layer to the ith node of the hidden layer, the output of the hidden layer is
Z＝Wx
Substituting it into Sigmoid function to get:
Figure BDA0003400790260000181
after hidden layer input is obtained, forward propagation is continued, the weight of the second layer neural network is in a random initialization mode, and the weight is expressed as:
V＝(v0，v1，v2，…，vn)
wherein viRepresenting the ith weight from the hidden layer node to the output layer node; the final output is then:
Figure BDA0003400790260000182
after obtaining the predicted value ^ y, defining a loss function by using MSE, wherein the loss function is as follows:
Figure BDA0003400790260000183
wherein m represents the total number of samples; the gradient descent is carried out on the neural network through back propagation, and the weight of the first layer of neural network is adjusted to be as follows:
Figure BDA0003400790260000184
where eta is learning rate, partial differential
Figure BDA0003400790260000185
Comprises the following steps:
Figure BDA0003400790260000186
and adjusting the weight of the first layer of neural network, and obtaining the weight according to a chain rule:
Figure BDA0003400790260000187
wherein ,
Figure BDA0003400790260000188
comprises the following steps:
Figure BDA0003400790260000189
after updating, completing a first round of circulation; the above circulation is repeatedly carried out until the iteration change of two times tends to 0, the iteration is stopped, and the model training is finished; for the next shift, after the staff schedules, new data is continuously put into the original model for training, and the model is continuously optimized;
after the training of the model is completed, when the shift is scheduled, the characteristics of the caregiver and the characteristics of the cared person are preprocessed, then the trained model is imported, and the caregiver score is obtained:
Scorei＝f(Xi，Xj)
wherein XiFeature vector, X, representing a caregiverjRepresenting the caregiver feature vector, f is the trained model function, ScoreiA score for i caregiver; and finally, selecting the caregiver with the highest score from the selectable caregivers. After the system is used, the station leader does not need to perform manual screening, and only needs to perform scheduling according to recommended nursing staff. The accuracy rate can reach 80 percent at present, and is gradually increased along with the increase of the shift scheduling data volume.
The embodiment of the invention also provides a caregiver intelligent recommendation scheduling method, which comprises the following steps: the information extraction module extracts the caregiver characteristics of the caregiver and the caretaker characteristics of the caretaker; the neural network building block performs the following actions: constructing a neural network according to the characteristics of the caregiver and the characteristics of the cared person; and training historical data of the caregiver characteristics and the caregivers characteristics to derive recommendation scores for pairwise pairings of the caregivers characteristics and the caregiver characteristics, wherein the recommendation scores characterize the degree to which the corresponding caregivers should be assigned to the corresponding caregivers; a scheduling module configured to perform the following actions: ranking the caregiver according to the recommendation score; and scheduling the caregiver to attend to the cared-giver according to the ranking.
The caregivers spend most of time on the path due to the fact that the two service places are far away from each other, and the caregivers have a large amount of time for waiting for the service time to start at the gate of the old people due to the fact that the two service times are far away from each other, so that the existing caregivers are unreasonable in work allocation, and the influence caused by shortage of the caregivers is more serious.
Based on the above insights, the invention provides an intelligent caregiver recommendation scheduling system, which establishes a mathematical model by extracting characteristics of a caregiver, characteristics of the old (characteristics of a nursed person), existing working characteristics and future working characteristics, through data analysis and machine learning, the most suitable scheduling scheme function is learned, and finally, the idle personnel are scheduled for future work according to the model, in addition, a neural network is constructed according to the characteristics of the nursing staff and the characteristics of the old people, historical data is put into training, a score between 0 and 1 is finally obtained, the score represents the degree of whether the caregiver should be assigned to the elderly, the qualified caregiver is ranked for score, recommended to the shift clerk, the shift clerk preferably selects the score to be the highest, the optimal solution of the nursing staff work allocation can be realized, and the defects caused by the short term of the nursing staff are greatly alleviated.
In summary, the above embodiments describe in detail different configurations of the caregiver intelligent recommendation scheduling system, and it goes without saying that the present invention includes but is not limited to the configurations listed in the above embodiments, and any modifications based on the configurations provided by the above embodiments are within the scope of the present invention. One skilled in the art can take the contents of the above embodiments to take a counter-measure.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The above description is only for the purpose of describing the preferred embodiments of the present invention, and is not intended to limit the scope of the present invention, and any variations and modifications made by those skilled in the art based on the above disclosure are within the scope of the appended claims.

Claims (11)

1. A caregiver intelligent recommendation scheduling system, comprising:
an information extraction module configured to extract caregiver features of a caregiver and caregiver-addressed features of a caregiver;
a neural network building module configured to perform the following actions:
constructing a neural network according to the characteristics of the caregiver and the characteristics of the cared person; and
training historical data of caregiver characteristics and caregiver characteristics to derive recommendation scores for pairwise caregiver characteristics and caregiver characteristics, wherein the recommendation scores characterize the degree to which a corresponding caregiver should be assigned to a corresponding caregiver;
a scheduling module configured to perform the following actions:
ranking the caregiver according to the recommendation score; and
and scheduling the nursing staff according to the ranking to nurse the cared person.
2. The caregiver intelligent recommendation scheduling system of claim 1 wherein:
the information extraction module is further configured to extract existing and future work features of the caregiver; and/or
The neural network building module comprises a mathematical model building module, and the mathematical model building module is configured to obtain the most appropriate scheduling scheme function through data analysis and machine learning according to the existing working characteristics and the future working characteristics; and/or
The scheduling module is further configured to perform the following actions:
scheduling future work for the idle personnel according to the scheduling scheme function; and
and obtaining qualified nursing staff according to the schedule of the future work.
3. The caregiver intelligent recommendation scheduling system of claim 2 wherein caregiver characteristics include one or more of: caregiver gender, caregiver age, caregiver category, caregiver work experience, company work experience, caregiver staff star, caregiver staff rating, number of complaints the caregiver has, total caregiver order, caregiver maximum scholastic calendar, caregiver political aspect, caregiver job type, caregiver marital status, caregiver height, caregiver weight, caregiver current order, caregiver last order time, caregiver next start time, caregiver last distance, caregiver next distance.
4. The caregiver intelligent recommendation scheduling system of claim 3 wherein caregiver-addressed features include one or more of: the sex of the cared, the age of the cared, the length of time the cared has signed up, the health grade of the cared, the average value of the grades of the cared, the number of complaints of the cared, the number of times the cared is served by the current cared and the number of times the cared is served by the current cared in 30 days.
5. The caregiver intelligent recommendation scheduling system of claim 4 further comprising:
the preprocessing module is configured to input the preprocessed caregiver characteristics and/or the preprocessed caregiver characteristics into the neural network construction module; the pretreatment comprises the following steps:
the sex of the caregiver and the sex of the caretaker are G e { M, F }, and specific values of sex input are calculated:
Figure FDA0003400790250000021
Ni＝1-M
wherein M represents allN represents a sheet of female staff, SiA score representing the ith work order;
preprocessing the gender input by calculating the specific value of the gender input, and distributing a caregiver with proper gender to a cared person;
inputting the characteristics of the caregiver class according to the requirements of the carereceiver:
Figure FDA0003400790250000022
where r is the requirement of the care recipient, piIs a caregiver class;
the caregiver station types include full-time, part-time, practice and retirement, the caregiver station type assigned preferentially is full-time, and the average single formula for each caregiver station type is:
Figure FDA0003400790250000023
wherein μiAverage unit amount, count (S) representing the ith caregiver typei) Represents the number of employees of the ith type; order (S)ij) Representing the order quantity of the jth caregiver of the ith type caregiver; after the average single amount of the caregivers of each post type is calculated, the average single amount is substituted into a softmax function, the softmax function is standardized, and meanwhile, the weight value of the dominant post is amplified:
Figure FDA0003400790250000031
wherein ,SiRepresenting the characteristic input of the i-th type of caregiver.
6. The caregiver intelligent recommendation scheduling system of claim 5 wherein said preprocessing further comprises:
calculating the time difference between the next order according to the previous single-ending time and the next single-starting time:
Figure FDA0003400790250000032
wherein TeIndicates the end time of the previous order, TsIndicates the start time, T, of the next orderiRepresenting a threshold time difference, if the time difference between the upper bill and the lower bill exceeds the threshold time difference, considering that a new bill can be inserted between the upper bill and the lower bill, and defining the threshold time difference according to the service condition;
Wtrepresenting the expected time of flight, PLNA feature input representing a time from a previous single bundle or a feature input representing a next single start time;
calculating the predicted journey time of the caregiver according to the predicted journey time from the previous order to the current order and the predicted journey time from the current order to the next order:
Figure FDA0003400790250000033
wherein (xi，yi) and (xj，yj) Respectively, longitude and latitude coordinate values of two places, v is the estimated speed of a vehicle used by a caregiver, and ρ is time redundancy which is preset according to the actual condition of the caregiver; the normalization of the time required for the journey is normalized using the modified mean variance:
Figure FDA0003400790250000034
where X is the time required for the trip, μ is the average of the caregiver's trip times, and σ is the variance of the caregiver's trip times.
7. The caregiver intelligent recommendation scheduling system of claim 6 wherein the neural network building module is further configured to:
constructing a BP (back propagation) neural network according to the preprocessed input characteristics of the caregiver and the characteristics of the caregivers, wherein the number of hidden layers of the BP neural network is one layer, an output layer of the BP neural network is a single neuron, and both the hidden layers and the output layer of the BP neural network adopt a Sigmoid function as an activation function;
after a BP neural network is constructed, a weight is given to a first layer neural network through an initialization weight module;
the initialization weight module adopts an expert system to calculate the weight of the first layer neural network, and the expert system adopts a large-scale multi-expert cooperation system, multi-knowledge representation, an integrated knowledge base, a self-organization problem solving mechanism, multi-subject cooperation problem solving and parallel reasoning, an expert system tool and environment and an artificial neural network knowledge acquisition and learning mechanism so as to have the weight of the first layer neural network with multiple knowledge bases and multiple main bodies.
8. The caregiver intelligent recommendation scheduling system of claim 7 wherein the weights for the first layer neural network comprise:
the highest scholastic calendar of the caregiver, the political face of the caregiver, the marital status of the caregiver, the height of the caregiver, the weight of the caregiver, the gender of the caregiver and the gender of the cared person are respectively 0.1;
the weight value of the health grade of the nursed person is 0.2;
the type of the caregiver, the type of the caregiver's posts, the working experience of the caregiver, and the weight of the signed time length of the cared person are 0.3;
the weight of the working experience of the company is 0.4;
the star rating of the nursing staff, the number of complaints of the nursing staff, the total single amount of the nursing staff, the age of the nursed person, the average value of the ratings of the nursed person and the number of complaints of the nursed person are 0.5;
the weight of the daily sheet amount of the nursing staff is 0.6;
the last single-ending time of the caregiver, the next starting time of the caregiver and the number of times of the current caregiver service of the cared person for the last 30 days are weighted to be 0.8;
the weight of the previous single distance of the caregiver, the next single distance of the caregiver, and the number of times the cared person is served by the current caregiver is 0.7.
9. The caregiver intelligent recommendation scheduling system of claim 8, wherein historical data is collected, historical scheduling data is input into the neural network for learning to obtain a final model for subsequent scheduling, and after the staff schedule, new data is continuously input into the original model for training to continuously optimize the model;
when the station leader needs to arrange nurses for the nursed persons, the nursing persons meeting the conditions are inquired, the data of the nursing persons and the nursed persons meeting the conditions form a vector to be used as the input of a model, the obtained output is sequenced and recommended to the station leader, and the station leader selects the first nursing person for scheduling;
preprocessing the input characteristics of the nursed person, wherein the weight calculated by the initialization module is obtained by carrying out pre-judgment by an expert system according to experience, so that the iteration times of training are reduced, and the model training speed is accelerated;
after characteristics of a caregiver and a nursed person are collected, a BP neural network is constructed, the number of hidden layers is one, and an output layer is a single neuron; the hidden layer and the output layer both adopt Sigmoid functions as activation functions.
10. The caregiver intelligent recommendation scheduling system of claim 9 wherein the training process for the model includes:
the feature input vector is:
x＝(x0，x1，x2，···，xp，xp+1，·，xn)T
wherein x1To xpCharacteristic of the caregiver, xp+1To xnIs a characteristic of the person to be cared for, x0Is a bias node; the weights of the first layer neural network are:
Figure FDA0003400790250000051
wherein wijRepresenting the weight from the jth node of the input layer to the ith node of the hidden layer, the output of the hidden layer is
Z＝Wx
Substituting it into Sigmoid function to get:
Figure FDA0003400790250000052
after hidden layer input is obtained, forward propagation is continued, the weight of the second layer neural network is in a random initialization mode, and the weight is expressed as:
V＝(υ0，υ1，υ2，…，υn)
wherein viRepresenting the ith weight from the hidden layer node to the output layer node; the final output is then:
Figure FDA0003400790250000061
after obtaining the predicted value ^ y, defining a loss function by using MSE, wherein the loss function is as follows:
Figure FDA0003400790250000062
wherein m represents the total number of samples; the gradient descent is carried out on the neural network through back propagation, and the weight of the first layer of neural network is adjusted to be as follows:
Figure FDA0003400790250000063
where eta is learning rate, partial differential
Figure FDA0003400790250000064
Comprises the following steps:
Figure FDA0003400790250000065
and adjusting the weight of the first layer of neural network, and obtaining the weight according to a chain rule:
Figure FDA0003400790250000066
wherein ,
Figure FDA0003400790250000067
comprises the following steps:
Figure FDA0003400790250000068
after updating, completing a first round of circulation; the above circulation is repeatedly carried out until the iteration change of two times tends to 0, the iteration is stopped, and the model training is finished; for the next shift, after the staff schedules, new data is continuously put into the original model for training, and the model is continuously optimized;
after the training of the model is completed, when the shift is scheduled, the characteristics of the caregiver and the characteristics of the cared person are preprocessed, then the trained model is imported, and the caregiver score is obtained:
Scorei＝f(Xi，Xj)
wherein XiFeature vector, X, representing a caregiverjRepresenting the caregiver feature vector, f is the trained model function, ScoreiA score for i caregiver; and finally, selecting the caregiver with the highest score from the selectable caregivers.
11. A caregiver intelligent recommendation scheduling method, comprising:
the information extraction module extracts the caregiver characteristics of the caregiver and the caretaker characteristics of the caretaker;
the neural network building block performs the following actions:
constructing a neural network according to the characteristics of the caregiver and the characteristics of the cared person; and
training historical data of caregiver characteristics and caregiver characteristics to derive recommendation scores for pairwise caregiver characteristics and caregiver characteristics, wherein the recommendation scores characterize the degree to which a corresponding caregiver should be assigned to a corresponding caregiver;
a scheduling module configured to perform the following actions:
ranking the caregiver according to the recommendation score; and
and scheduling the nursing staff according to the ranking to nurse the cared person.
CN202111499182.8A 2021-12-09 2021-12-09 Intelligent recommendation scheduling system for nursing staff Active CN114117238B (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202111499182.8A CN114117238B (en) 2021-12-09 2021-12-09 Intelligent recommendation scheduling system for nursing staff
CN202311320590.1A CN117393124A (en) 2021-12-09 2021-12-09 Intelligent scheduling method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111499182.8A CN114117238B (en) 2021-12-09 2021-12-09 Intelligent recommendation scheduling system for nursing staff

Related Child Applications (1)

Application Number Title Priority Date Filing Date
CN202311320590.1A Division CN117393124A (en) 2021-12-09 2021-12-09 Intelligent scheduling method

Publications (2)

Publication Number Publication Date
CN114117238A true CN114117238A (en) 2022-03-01
CN114117238B CN114117238B (en) 2023-08-29

Family

ID=80363801

Family Applications (2)

Application Number Title Priority Date Filing Date
CN202111499182.8A Active CN114117238B (en) 2021-12-09 2021-12-09 Intelligent recommendation scheduling system for nursing staff
CN202311320590.1A Pending CN117393124A (en) 2021-12-09 2021-12-09 Intelligent scheduling method

Family Applications After (1)

Application Number Title Priority Date Filing Date
CN202311320590.1A Pending CN117393124A (en) 2021-12-09 2021-12-09 Intelligent scheduling method

Country Status (1)

Country Link
CN (2) CN114117238B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116612870A (en) * 2023-07-17 2023-08-18 山东圣剑医学研究有限公司 General surgery patient data management method

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150154721A1 (en) * 2013-12-02 2015-06-04 Talksession, Inc. System, apparatus and method for user to obtain service from professional
CN107111808A (en) * 2015-01-07 2017-08-29 皇家飞利浦有限公司 Pair with object interact progress scheduling
US20170293878A1 (en) * 2016-04-12 2017-10-12 Softvu Llc System and process for matching seniors and staffers with senior living communities
US20200388360A1 (en) * 2014-12-10 2020-12-10 Koninklijke Philips N.V. Methods and systems for using artificial neural networks to generate recommendations for integrated medical and social services
CN112288394A (en) * 2020-10-29 2021-01-29 中国民用航空总局第二研究所 System for verifying scheduling test of controller
CN112837774A (en) * 2021-02-09 2021-05-25 福寿康(上海)医疗养老服务有限公司 Intelligent assessment method for home-based care service for aged people
CN113470796A (en) * 2021-06-08 2021-10-01 华中科技大学同济医学院附属协和医院 Nursing scheduling management system and using method thereof
CN113486073A (en) * 2021-05-08 2021-10-08 华东师范大学 Aged care caregiver care information collaborative optimization method oriented to digital environment
CN113643798A (en) * 2021-08-30 2021-11-12 平安医疗健康管理股份有限公司 Method and device for matching caregivers for disabled persons and computer equipment

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150154721A1 (en) * 2013-12-02 2015-06-04 Talksession, Inc. System, apparatus and method for user to obtain service from professional
US20200388360A1 (en) * 2014-12-10 2020-12-10 Koninklijke Philips N.V. Methods and systems for using artificial neural networks to generate recommendations for integrated medical and social services
CN107111808A (en) * 2015-01-07 2017-08-29 皇家飞利浦有限公司 Pair with object interact progress scheduling
US20170293878A1 (en) * 2016-04-12 2017-10-12 Softvu Llc System and process for matching seniors and staffers with senior living communities
CN112288394A (en) * 2020-10-29 2021-01-29 中国民用航空总局第二研究所 System for verifying scheduling test of controller
CN112837774A (en) * 2021-02-09 2021-05-25 福寿康(上海)医疗养老服务有限公司 Intelligent assessment method for home-based care service for aged people
CN113486073A (en) * 2021-05-08 2021-10-08 华东师范大学 Aged care caregiver care information collaborative optimization method oriented to digital environment
CN113470796A (en) * 2021-06-08 2021-10-01 华中科技大学同济医学院附属协和医院 Nursing scheduling management system and using method thereof
CN113643798A (en) * 2021-08-30 2021-11-12 平安医疗健康管理股份有限公司 Method and device for matching caregivers for disabled persons and computer equipment

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
E.NAUDIN 等: "Analysis of three mathematical models of the staff rostering problem", pages 23 - 38 *
张志云 等: "护士自助排班的临床实践", vol. 50, no. 11, pages 1326 - 1330 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116612870A (en) * 2023-07-17 2023-08-18 山东圣剑医学研究有限公司 General surgery patient data management method
CN116612870B (en) * 2023-07-17 2023-10-10 山东圣剑医学研究有限公司 General surgery patient data management method

Also Published As

Publication number Publication date
CN117393124A (en) 2024-01-12
CN114117238B (en) 2023-08-29

Similar Documents

Publication Publication Date Title
Hofstede et al. Measuring organizational cultures: A qualitative and quantitative study across twenty cases
Ceder Public transit planning and operation: Modeling, practice and behavior
Armistead Service operations strategy: framework for matching the service operations task and the service delivery system
Trappe et al. Occupational sex segregation and family formation in the former East and West Germany
Shahwahid et al. Assessing service quality of community-based ecotourism: A case study from Kampung Kuantan Firefly Park
Isikli et al. Identifying key factors of rail transit service quality: an empirical analysis for Istanbul
CN114117238A (en) Intelligent caregiver recommendation scheduling system
Rabren et al. Postschool outcome data collection programs: Examples from two states
Franco et al. Achieving quality through problem solving and process improvement
Chou et al. Longitudinal assessment of high-speed rail service delivery, satisfaction and operations: A study of Taiwan and Korea systems
Ahmad et al. Hotel Employees’ Burnout and Turnover Intentions
Dieleman et al. A three-step framework for capacity planning in a nursing home context
Abraham et al. Leadership Competencies and Employee Satisfaction in Nursing Homes.
CN115829214A (en) Scheduling method and scheduling system
Ditmore et al. A study of day care's effect on the labor force participation of low-income mothers
Ongy Analysis of college enrollment system using Six-Sigma (DMAIC) methodology
Litwin et al. Employing disabled workers in production: simulating the impact on performance and service level
Xiao Equilibrium Sorting and the Gender Wage Gap
Tansitpong Unveiling Key Differentiated Service Dimensions of the Airlines Performances in the COVID-19 Aftermath: An Empirical Investigation on Operational and Service Outcomes on Market Share and Load Factor
McGuinness et al. EVALUATION OF PLC PROGRAMME PROVISION. ESRI RESEARCH SERIES NUMBER 61 JANUARY 2018
Ould Sidi et al. A novel approach to developing and evaluating regulation strategies for urban transport disrupted networks
Murad Revolutionizing Tourism Performance in Malaysia through Artificial Intelligence: Optimizing Flight Booking and Accommodation for Elevated Visitor Experiences
Lin Psychological health correlation of express delivery workers’ occupational stress in the information logistics environment
Manjunatha Working environment of lady veterinary surgeons of animal husbandry department of Kerala state
Zar CUSTOMER SATISFACTION ON WAITING LINE MANAGEMENT AT MYANMA FOREIGN TRADE BANK

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