Guest room cleaning sequence generation method and system and medium

Abstract

The invention relates to a cleaning sequence generation method and system and a medium, and relates to the field of hotel systems.The method comprises the steps that state information of all guest rooms is obtained, and the state information comprises position information, check-in information, reservation information and customer demand information; forming a training sample based on the historical guest room state information and the historical cleaning priority data, and training a preset initial model to obtain a multi-layer sensor; and processing the state information of each guest room based on the multi-layer sensor, and outputting the guest room cleaning sequence. According to the invention, the automatic decision-making of the cleaning priority in the hotel management process is realized, and the intelligent degree is greatly improved.

Description

Technical Field

[0001] The invention relates to the field of hotel systems, and in particular to a guest room cleaning sequence generation method, system and medium. Background Art

[0002] Room cleaning is an essential part of the hotel's daily operations. Its main purpose is to ensure that guests can enjoy a clean and comfortable environment during their stay. Effective cleaning arrangements can not only improve the hotel's operational efficiency, but also directly affect guests' satisfaction and overall experience.

[0003] Traditional housekeeping methods often rely on manual judgment and paper records. Cleaners clean rooms one by one according to a checklist or instructions from management, and the order and priority of cleaning are usually determined by experience and intuition. Although this method is simple, it may be difficult to handle a large number of rooms or special needs.

[0004] This traditional method has many problems. Due to its reliance on manual judgment, it is easy to cause an unreasonable cleaning order or omission of important rooms, affecting overall efficiency. At the same time, paper records are prone to errors or omissions and are difficult to update in a timely manner, resulting in the inability to adjust cleaning tasks in real time and the inability to flexibly respond to emergencies, which in turn reduces guest satisfaction. Summary of the invention

[0005] In view of the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a guest room cleaning order generation method, system and medium to achieve automated decision-making on cleaning priorities, thereby greatly improving the degree of intelligence.

[0006] To achieve the above purpose, the present invention adopts the following technical solution.

[0007] In a first aspect, the present invention provides a guest room cleaning sequence generation method, system and medium, which adopt the following technical solutions: Acquire status information of each guest room, including location information, check-in information, reservation information, and customer demand information; Based on the historical guest room status information and historical cleaning priority data, the training samples are formed and the preset initial model is trained to obtain a multi-layer perceptron; The state information of each guest room is processed based on the multi-layer perceptron, and the guest room cleaning sequence is output.

[0008] Furthermore, in the above-mentioned guest room cleaning sequence generation method, the state information of each guest room is obtained, and the state information includes location information, check-in information, reservation information and customer demand information, including: Get the specific location coordinates of each room on the floor and convert them into corresponding location numerical features; Get the current occupancy status of each room and convert it into the corresponding status type variable; Get the reservation status of each vacant room and convert it into the corresponding reservation type variable; Get the variable indicating whether there is a cleaning requirement in each guest room and convert it into a corresponding requirement Boolean variable; The position numerical feature, the state type variable, the predetermined type variable and the required Boolean variable are stored as preset structure data to form the state information.

[0009] Furthermore, in the above-mentioned guest room cleaning sequence generation method, the training samples are formed based on the historical guest room status information and the historical cleaning priority data, and the preset initial model is trained to obtain a multilayer perceptron, including: Extract historical guest room status information and match it with corresponding historical cleaning priority data to obtain a training set; Performing standardization processing on the position numerical features corresponding to the position information; Performing one-hot encoding on the state type variables, reservation type variables, and demand Boolean variables corresponding to the check-in information, reservation information, and customer demand information; Based on a preset loss function, the initial model is trained using the training set, the weights are adjusted through a back-propagation algorithm, and the loss function is minimized to obtain a final multi-layer perceptron.

[0010] Furthermore, in the above-mentioned guest room cleaning sequence generation method, the step of processing the status information of each guest room based on the multi-layer perceptron and outputting the guest room cleaning sequence includes: Input the currently acquired guest room status information into the trained multi-layer perceptron; Calculating the cleaning priority level of the corresponding guest room in the multi-layer perceptron; Based on the cleaning priority levels, all rooms are sorted to generate a priority list of cleaning tasks; The cleaning priority level is set as a discrete category label or a continuous priority score.

[0011] Furthermore, in the above-mentioned guest room cleaning order generation method, when the cleaning priority level is set to a discrete category label, the loss function is set to a cross entropy loss function; when the cleaning priority level is set to a continuous priority score, the loss function is set to a mean square error loss function.

[0012] Furthermore, in the above-mentioned guest room cleaning order generation method, when the cleaning priority level is set as a discrete category label, if there are rooms with the same label, the rooms with the same label are clustered, and the priority order of the rooms with the same label is output based on the clustering result.

[0013] Furthermore, the above-mentioned guest room cleaning sequence generation method further includes: The guest room cleaning sequence is sent to a preset mobile terminal, and the mobile terminal is used to assist cleaning personnel in performing guest room cleaning operations based on the guest room cleaning sequence.

[0014] In a second aspect, the present invention provides a guest room cleaning sequence generation system, which adopts the following technical solution: A status acquisition module, at least used to acquire status information of each guest room, wherein the status information includes location information, check-in information, reservation information and customer demand information; A model training module, at least used to form a training sample based on historical guest room status information and historical cleaning priority data, and train a preset initial model to obtain a multi-layer perceptron; The sequence generation module is at least used to process the status information of each guest room based on the multi-layer perceptron and output the guest room cleaning sequence.

[0015] Furthermore, the above-mentioned guest room cleaning sequence generation system further includes: The mobile terminal is at least used to obtain the room cleaning sequence generated by the sequence generation module.

[0016] In a third aspect, the present invention provides a readable storage medium, which adopts the following technical solution. The readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, a method for generating a guest room cleaning order as described in any one of the first aspects above is implemented.

[0017] In summary, compared with the prior art, the present invention includes at least one of the following beneficial technical effects: The present invention combines historical data and a multi-layer perceptron model to perform intelligent analysis on the status information of each guest room, thereby realizing the automatic generation of the guest room cleaning sequence. This not only improves the efficiency and accuracy of the cleaning task, but can also dynamically adapt to different guest room status and customer needs, thereby optimizing hotel operations and improving customer satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 It is a flowchart of a specific embodiment of a guest room cleaning sequence generation method of the present invention.

[0020] Figure 2It is a flowchart of another specific embodiment of a guest room cleaning sequence generating method of the present invention.

[0021] Figure 3 It is a flowchart of another specific embodiment of a guest room cleaning sequence generating method of the present invention.

[0022] Figure 4 It is a structural schematic diagram of a specific embodiment of a guest room cleaning sequence generation system of the present invention. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application. In addition, it should be understood that the specific implementation methods described herein are only used to illustrate and explain the present application, and are not used to limit the present application.

[0024] It should be noted that the description order of the following embodiments is not intended to limit the preferred order of the embodiments of the present application. In addition, in the following embodiments, the description of each embodiment has its own emphasis, and for parts not described in detail in one embodiment, reference can be made to the relevant description of other embodiments.

[0025] The method steps described in the embodiments of the present invention may be executed in the order described in the specific implementation manner, or the execution order of each step may be adjusted according to actual needs and on the premise that the technical problem can be solved, which will not be listed here one by one.

[0026] Reference Figure 1 , a guest room cleaning sequence generation method, comprising: S1, obtaining status information of each guest room, wherein the status information includes location information, check-in information, reservation information and customer demand information; S2, based on the historical guest room status information and historical cleaning priority data, the training samples are trained to obtain a multi-layer perceptron; S3, processing the status information of each guest room based on the multi-layer perceptron, and outputting the guest room cleaning order.

[0027] Specifically, step S1 obtains the status information of each guest room from the hotel management system (PMS) in real time, including the room's location information, current check-in status (such as checked-in, checked-out), reservation status (such as upcoming reservation information), and special customer needs (such as VIP service requests, do not disturb status, etc.), and formats and encodes this information to provide basic data input for subsequent model processing and cleaning sequence generation.

[0028] Step S2 collects the status information of each guest room (including location information, check-in information, reservation information and customer demand information) and the corresponding historical cleaning priority from the historical data to construct a complete training data set. Then, the data is preprocessed, including standardizing the numerical features, encoding the category features, processing missing values, etc. Next, the preprocessed data is divided into a training set, a validation set and a test set, and the training set is used to train the preset multi-layer perceptron model, and the model parameters are adjusted to minimize the loss function. Through the evaluation on the validation set, the model performance is continuously optimized, and finally the accuracy of the model is verified on the test set to ensure that the multi-layer perceptron model can accurately output the cleaning priority according to the current guest room status information.

[0029] Step S3 inputs the status information (including location information, check-in information, reservation information and customer demand information) of each guest room obtained from the current hotel management system into the trained multi-layer perceptron model. The model will calculate based on these input features and output the cleaning priority score of each room. Then, all rooms are sorted according to these scores to generate a final cleaning order list. If there are multiple rooms with the same priority score, these rooms can be further subdivided and sorted by a clustering algorithm to optimize the cleaning path and order, and the generated cleaning order list will eventually be fed back to the cleaning team for execution.

[0030] Based on the above embodiment, through intelligent means, a multi-layer perceptron model is used to combine historical data and real-time status information to automatically generate an optimized room cleaning sequence, which significantly improves the efficiency and accuracy of the cleaning task. It can dynamically adapt to the real-time status and customer needs of different rooms, thereby optimizing the hotel operation process and improving guests' stay experience and satisfaction.

[0031] Further, as an implementation mode of the present invention, step S1, obtaining status information of each guest room, the status information including location information, check-in information, reservation information and customer demand information, includes: Get the specific location coordinates of each room on the floor and convert them into corresponding location numerical features; Get the current occupancy status of each room and convert it into the corresponding status type variable; Get the reservation status of each vacant room and convert it into the corresponding reservation type variable; Get the variable indicating whether there is a cleaning requirement in each guest room and convert it into a corresponding requirement Boolean variable; The position numerical feature, the state type variable, the predetermined type variable and the required Boolean variable are stored as preset structure data to form the state information.

[0032] For example, assume that the hotel has three rooms: 101, 102, and 103. It is necessary to obtain the status information of these three rooms, including location information, check-in information, reservation information, and customer demand information, and convert this information into corresponding numerical or category features for subsequent processing.

[0033] Get the position information and convert it into numerical features: Room 101: Located on the 1st floor, X coordinate is 15, Y coordinate is 30; Room 102: Located on the 1st floor, X coordinate is 35, Y coordinate is 40; Room 103: Located on the 1st floor, X coordinate is 25, Y coordinate is 20. If converted into numerical features, the X and Y coordinates of the room can be directly used as the position numerical features. For example, the position numerical feature of room 101 can be expressed as (15, 30).

[0034] Get the check-in status and convert it into a status type variable: Room 101: Currently occupied, status is "Occupied"; Room 102: Checked out, status is "Vacant"; Room 103: Vacant but reserved, status is "Reserved". Convert to Status Type Variable You can convert the status to a numeric or categorical variable. For example, "Occupied" is coded as 1, "Vacant" is coded as 2, and "Reserved" is coded as 3.

[0035] Get the booking status and convert it into a booking type variable: Room 101: No reservation, reservation status is "None"; Room 102: No reservation, reservation status is "None"; Room 103: There is an urgent reservation, and the reservation status is "Urgent". Converting to a booking type variable converts the booking status into a categorical variable. For example, "None" (no booking) is coded as 0, and "Urgent" (urgent booking) is coded as 1.

[0036] Get the cleaning requirement and convert it into a requirement Boolean variable: Room 101: No special cleaning requirements, the requirement status is "False"; Room 102: No special cleaning requirements, the requirement status is "False"; Room 103: There is a cleaning requirement, and the requirement status is "True". Convert to Requirement Boolean Variable allows you to convert the requirement status to a Boolean variable. For example, "False" is encoded as 0 and "True" is encoded as 1.

[0037] The numerical features, status type variables, booking type variables, and demand Boolean variables for each room are stored as structured data rows. For example, a row in the structured data table can be represented as follows: Room 101: (15,30),1,0,0 Room 102: (35,40),2,0,0 Room 103: (25,20),3,1,1.

[0038] Through these steps, S1 converts the acquired location information, check-in status, reservation information, and cleaning requirements into numerical features, status type variables, reservation type variables, and requirement Boolean variables, and stores these features as structured data rows. These data will be used for subsequent model processing and cleaning order generation.

[0039] Further, see Figure 2 As an implementation of the present invention, step S2, based on the historical guest room status information and the historical cleaning priority data to form a training sample, train the preset initial model to obtain a multilayer perceptron, including: S21, extracting historical guest room status information and matching it with corresponding historical cleaning priority data to obtain a training set; S22, performing standardization processing on the position numerical feature corresponding to the position information; S23, performing one-hot encoding on the state type variables, reservation type variables, and demand Boolean variables corresponding to the check-in information, reservation information, and customer demand information; S24, based on a preset loss function, using the training set to train the initial model, adjusting the weights through a back propagation algorithm, minimizing the loss function to obtain a final multilayer perceptron.

[0040] Specifically, step S21 extracts a complete data set including historical status information such as room location information, occupancy status, reservation information, customer demand information, etc. from the historical database, and extracts historical cleaning priority labels corresponding to these status information. A training set is formed by matching the status information of each room with its corresponding cleaning priority one by one, wherein each training sample consists of a set of status information features (such as location, occupancy status, reservation status, customer demand) and a corresponding cleaning priority label, providing a data basis for subsequent model training.

[0041] Illustrative, historical data example: Room 101: Location information (15,30), occupancy status "Occupied" (1), reservation status "None" (0), customer demand "False" (0); Room 102: Location information (35, 40), occupancy status "Vacant" (2), reservation status "None" (0), customer demand "False" (0); Room 103: Location information (25, 20), occupancy status "Reserved" (3), reservation status "Urgent" (1), customer demand "True" (1).

[0042] Corresponding cleaning priority: Room 101: Priority 3; Room 102: Priority 2; Room 103: Priority 1.

[0043] The matching result is: match the guest room status information of each room with its historical cleaning priority, and obtain the training set: Training sample 1: [(15,30),1,0,0]->priority 3; Training sample 2: [(35,40),2,0,0]->priority 2; Training sample 3: [(25,20),3,1,1]->priority 1.

[0044] Specifically, in step S22, the mean and standard deviation of each room position information in the training set are first calculated, and then the X and Y coordinates in the position information are standardized to convert them into numerical features with a mean of 0 and a standard deviation of 1. This process can eliminate the dimensional differences in different room position data, making the model more stable and efficient in processing these position features, and providing standardized data input for subsequent model training.

[0045] For example, the normalization formula for normalization processing may be: Among them, X is the original eigenvalue, μ is the mean of the feature, and ρ is the standard deviation of the feature.

[0046] Calculate the mean and standard deviation: X coordinate mean = (15 + 35 + 25) / 3 = 25; Y coordinate mean = (30 + 40 + 20) / 3 = 30; X coordinate standard deviation = sqrt([(15-25)^2+(35-25)^2+(25-25)^2] / 3) = 10; Y coordinate standard deviation = sqrt([(30-30)^2+(40-30)^2+(20-30)^2] / 3)=10.

[0047] The standardization process is: Room 101 position (15,30) -> normalized position [(15-25) / 10, (30-30) / 10] = [-1.0, 0.0]; Room 102 position (35,40) -> normalized position [(35-25) / 10, (40-30) / 10] = [1.0, 1.0]; Room 103 position (25, 20) -> normalized position [(25-25) / 10, (20-30) / 10] = [0.0, -1.0].

[0048] Step S23 performs one-hot encoding on the categorical features such as the room's check-in information, reservation information, and customer demand information, converting each categorical variable into multiple binary features. For example, the check-in status "Occupied", "Vacant", and "Reserved" are encoded as [1,0,0], [0,1,0], and [0,0,1], respectively; the reservation status "None" and "Urgent" are encoded as [1,0] and [0,1]; the customer demand "True" and "False" are encoded as [0,1] and [1,0]. This encoding method converts the categorical variables into a numerical format that can be directly processed by the model, ensuring that each categorical feature is independently represented and processed in model training.

[0049] Exemplarily, the One-Hot Encoding process includes: Occupancy status code: ["Occupied", "Vacant", "Reserved"]->[1,0,0], [0,1,0], [0,0,1]; Booking status code: ["None", "Urgent"]->[1,0], [0,1]; Customer requirement encoding: ["False","True"]->[1,0], [0,1].

[0050] The encoded training samples are: Training sample 1: [-1.0, 0.0], [1, 0, 0], [1, 0], [1, 0] -> priority 3; Training sample 2: [1.0, 1.0], [0, 1, 0], [1, 0], [1, 0] -> priority 2; Training sample 3: [0.0,-1.0],[0,0,1],[0,1],[0,1]->priority 1.

[0051] The specific implementation of step S24 selects a suitable loss function (such as mean square error MSE), then inputs the standardized and one-hot encoded training set into the multi-layer perceptron model, calculates the predicted value of each training sample through forward propagation, and calculates the loss based on the error between the predicted value and the actual cleaning priority. Next, the gradient of the loss function to the model weight is calculated using the back propagation algorithm, and the model weight is adjusted using an optimizer (such as Adam or stochastic gradient descent method) to gradually minimize the loss function. Through multiple iterative training, an optimized multi-layer perceptron model is finally obtained, which can effectively predict the cleaning priority of the guest room.

[0052] Exemplarily, a multilayer perceptron model is selected for training. First, the model architecture of the multilayer perceptron model is initialized: input layer: 8 neurons, corresponding to 8 input features (2 standardized location features + 3 one-hot encoded check-in status features + 2 one-hot encoded booking status features + 1 one-hot encoded customer demand feature); Hidden layer: Assume that 2 hidden layers are used, each containing 10 neurons; Output layer: 1 neuron, used to output the predicted cleaning priority score.

[0053] The mean square error (MSE) is selected as the loss function to measure the difference between the model's predicted cleaning priority and the actual cleaning priority. The MSE formula is: in, is the cleaning priority predicted by the model, y i is the actual cleaning priority and n is the number of samples.

[0054] For each training sample, the input feature enters the model through the input layer, undergoes a nonlinear transformation (such as the ReLU activation function) in the hidden layer, and finally reaches the output layer to generate a predicted cleaning priority score. For example, suppose that for training sample 1, the model predicts a priority score of 2.8, while the actual priority is 3.

[0055] The MSE loss function is used to calculate the error between the model's predicted value and the actual value. For example, for training sample 1, the MSE loss is: L = (2.8-3) 2 = 0.04. The losses of all training samples are averaged to obtain the overall loss value.

[0056] Based on the calculated loss, the gradient of each model parameter (weight and bias) is calculated through the back-propagation algorithm. Use an optimizer (such as Adam or stochastic gradient descent) to adjust the model weights in the opposite direction of the gradient to minimize the loss function.

[0057] The weight update formula (taking SGD as an example) is: Where w is the weight, η is the learning rate, is the gradient of the loss function with respect to the weights.

[0058] Through multiple rounds of iterations (multiple epochs), the process of forward propagation, loss calculation, back propagation, and weight update is repeated to gradually reduce the model's loss until the model's performance on the training set reaches a satisfactory level. After multiple rounds of training, the model is able to better predict the cleaning priority. At this point, the final multi-layer perceptron model has been optimized and can be used to process actual data and output the priority order of room cleaning.

[0059] Through these steps, S24 realized the complete process from data input, model training to optimization, and finally obtained a multi-layer perceptron model that can accurately predict the cleaning priority based on the room status information.

[0060] Furthermore, as an implementation mode of the present invention, step S3, based on the multilayer perceptron processing the status information of each guest room, outputting the guest room cleaning order, includes: inputting the currently acquired guest room status information into the trained multilayer perceptron; calculating the cleaning priority level of the corresponding guest room in the multilayer perceptron; based on the cleaning priority level, sorting all rooms to generate a priority list of cleaning tasks; wherein the cleaning priority level is set as a discrete category label or a continuous priority score.

[0061] Specifically, the status information of each guest room currently obtained in real time (including features such as location information, check-in information, reservation information, and customer demand information) is standardized and one-hot encoded, and then input into the trained multi-layer perceptron model. The model calculates the cleaning priority level of each room through forward propagation. Then, all rooms are sorted according to the calculated cleaning priority level, and the room with the smaller priority value has higher priority. Finally, these sorting results are used to generate a priority list of cleaning tasks, which is fed back to the cleaning team to arrange the cleaning work for the day, thereby ensuring that the cleaning tasks are carried out in the optimal order.

[0062] Furthermore, the cleaning priority output by the model can be expressed in two forms. Discrete category labels indicate that the priority is divided into several fixed categories (such as 1, 2, 3, 4, 5), and each category corresponds to a different cleaning priority order; continuous priority scoring represents the priority as a continuous numerical range (such as any floating point number between 1.0 and 5.0). This score can more finely reflect the cleaning priority of each room, allowing for more flexible and accurate sorting. Depending on the needs of the specific application scenario, you can choose to set the cleaning priority in a discrete or continuous manner to achieve the optimal cleaning task schedule.

[0063] Furthermore, as an embodiment of the present invention, when the cleaning priority level is set to a discrete category label, the loss function is set to a cross entropy loss function; when the cleaning priority level is set to a continuous priority score, the loss function is set to a mean square error loss function.

[0064] Specifically, according to the different representation methods of the cleaning priority levels, a suitable loss function can be selected to optimize the model training process: When the cleaning priority is set to a fixed category (for example, priority 1, 2, 3, 4, 5), the output of the model is a classification result, which classifies the cleaning priority of each room into these discrete categories. The cross entropy loss function is used to measure the difference between the category probability distribution predicted by the model and the actual category in the classification problem. The cross entropy loss function calculates the logarithmic loss between the predicted probability and the true category to guide the model on how to adjust the weights to reduce misclassification and improve classification accuracy.

[0065] Exemplarily, the formula for the cross entropy loss function is: Among them, y i is the label of the actual category (represented by one-hot encoding), p i is the probability that the model predicts the category, and C is the number of categories.

[0066] When the cleaning priority is set to a continuous value (such as any floating point number between 1.0 and 5.0), the output of the model is a specific priority score that can more carefully reflect the priority differences between different rooms. The mean square error loss function is used to measure the difference between the model prediction value and the actual value in the regression problem. The mean square error loss function calculates the average of the square of the error between the predicted value and the true value, and guides the model to adjust the weight so that the predicted value is as close to the true value as possible, thereby improving the accuracy of the model. The mean square error loss function is described in the above embodiment and will not be repeated here.

[0067] By selecting an appropriate loss function, the model can more effectively learn and predict cleaning priorities, improving the rationality and efficiency of task scheduling.

[0068] Furthermore, as an embodiment of the present invention, when the cleaning priority level is set as a discrete category label, if there are rooms with the same label, the rooms with the same label are clustered, and the priority order of the rooms with the same label is output based on the clustering result.

[0069] Specifically, when the cleaning priority level is set as a discrete category label, if multiple rooms are assigned to the same priority label (for example, multiple rooms are marked as priority 2), the cleaning order of these rooms needs to be further subdivided. The specific implementation method is: first, use a clustering algorithm (such as K-Means clustering) to cluster these rooms with the same priority label, and divide them into several subgroups based on the room's location information, cleaning requirements, reservation status and other features. After clustering is completed, the rooms are sorted according to specific rules (such as distance, urgency of cleaning needs, etc.) in each subgroup to generate a detailed priority order. Ultimately, these subdivided cleaning orders will be used to optimize the cleaning arrangements of rooms with the same label, ensuring the optimal cleaning efficiency and resource allocation in rooms with the same priority label.

[0070] Further, see Figure 3 As an embodiment of the present invention, the guest room cleaning sequence generation method of the present invention further includes: S4: Sending the guest room cleaning sequence to a preset mobile terminal, where the mobile terminal is used to assist cleaning personnel in performing guest room cleaning operations based on the guest room cleaning sequence.

[0071] The generated room cleaning order is sent to the preset mobile terminal (such as a smartphone or tablet) used by the cleaning staff. The cleaning staff can view the cleaning priority list generated by the system in real time through the mobile terminal and perform cleaning operations in that order. This function is designed to improve the efficiency and accuracy of cleaning operations, allowing cleaning staff to complete tasks efficiently and orderly, and ensure optimal room cleaning arrangements and execution.

[0072] The embodiment of the invention also discloses a guest room cleaning sequence generating system.

[0073] Reference Figure 4 The room cleaning sequence generation system includes a state acquisition module 1, a model training module 2, a sequence generation module 3 and a mobile terminal 4.

[0074] The status acquisition module 1 is integrated with the hotel management system (such as the PMS system) to extract the status information of each guest room in real time, including the room's location information (such as floor and specific location coordinates), current check-in status (such as checked-in, vacant, checked-out, etc.), reservation information (such as whether there is an upcoming reservation, reservation type, etc.), and special needs of customers (such as VIP service requests, do not disturb status, additional cleaning requests, etc.). After preliminary processing and formatting, these data are stored as standardized numerical features and categorical variables to provide input data for subsequent model training and cleaning sequence generation. In addition, the module can also update the status information regularly or as needed to ensure the real-time and accuracy of the cleaning sequence.

[0075] Model training module 2 extracts a large amount of historical guest room status information (such as location information, occupancy status, reservation information, and customer demand information) and the corresponding historical cleaning priority data from the historical database to construct a complete training sample set. These sample data are preprocessed, including standardization of numerical features, unique hot encoding of categorical variables, etc., to ensure that the data format is uniform and suitable for model input. Use these preprocessed training data to train the preset initial model (such as a multilayer perceptron model), calculate the predicted value through forward propagation, and use the loss function (such as cross entropy loss or mean square error) to evaluate the prediction error of the model. The weight of the model is continuously adjusted through the back propagation algorithm, and the model is iteratively optimized until the loss function converges, and finally an optimized multilayer perceptron model is obtained, which can accurately predict the cleaning priority of the guest room based on the real-time status information.

[0076] The sequence generation module 3 inputs the status information of each guest room (including location information, occupancy status, reservation information and customer demand information) obtained in real time into the trained multi-layer perceptron model, and outputs the cleaning priority level of each room through the forward propagation calculation model. All rooms are sorted according to the priority level generated by the model, and the rooms with higher priority values ​​are cleaned first. If multiple rooms have the same priority label, the module can further subdivide and sort these rooms through a clustering algorithm, considering additional features such as room location or the urgency of cleaning needs, and optimizing the cleaning order. Finally, the module generates an ordered cleaning task list and sends it to the cleaning team to guide the actual cleaning work.

[0077] The mobile terminal 4 maintains real-time communication with the sequence generation module 3 through a wireless network or a local area network to obtain the latest generated room cleaning sequence list. The cleaning staff logs into the system through a mobile terminal (such as a smart phone, tablet computer or a dedicated handheld device) to view the cleaning sequence list in real time. The mobile terminal 4 will display detailed information of the cleaning task in a user-friendly interface, including room number, priority, specific cleaning requirements (such as VIP services or additional cleaning requests), etc. In addition, the mobile terminal 4 also has a feedback function. The cleaning staff can mark the completion status after completing each task, and the system will automatically update the task progress. This method not only improves the efficiency and accuracy of the cleaning operation, but also ensures that the cleaning staff performs tasks in the optimal order and improves the overall work coordination.

[0078] Based on this, the guest room cleaning sequence generation system collects and processes the guest room status information in real time through the status acquisition module 1, uses the model training module 2 to train and optimize the multilayer perceptron model based on historical data, and then the sequence generation module 3 generates the optimized cleaning task sequence according to the real-time data, and accurately conveys the cleaning instructions to the cleaning staff through the mobile terminal 4, finally realizing efficient and intelligent guest room cleaning task arrangement, improving cleaning efficiency, accuracy and guest satisfaction.

[0079] The embodiment of the present invention also discloses a readable storage medium.

[0080] A readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the guest room cleaning order generation method described in any one of the above embodiments. The computer-readable storage medium may include: any entity or device capable of carrying a computer program, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), and a software distribution medium, etc. The computer program includes a computer program code. The computer program code may be in source code form, object code form, an executable file, or some intermediate form, etc. The computer-readable storage medium may include: any entity or device capable of carrying a computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), and a software distribution medium, etc.

[0081] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0082] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, apparatus or device (such as a computer-based system, a system including a processing module, or other system that can fetch instructions from an instruction execution system, apparatus or device and execute instructions), or used in combination with these instruction execution systems, apparatuses or devices.

[0083] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for generating a guest room cleaning sequence, characterized in that: include: Acquire status information of each guest room, including location information, check-in information, reservation information, and customer demand information; Based on the historical guest room status information and historical cleaning priority data, the training samples are formed and the preset initial model is trained to obtain a multi-layer perceptron; The state information of each guest room is processed based on the multi-layer perceptron, and the guest room cleaning sequence is output.

2. The guest room cleaning sequence generation method according to claim 1, characterized in that: The state information of each guest room is obtained, and the state information includes location information, check-in information, reservation information and customer demand information, including: Get the specific location coordinates of each room on the floor and convert them into corresponding location numerical features; Get the current occupancy status of each room and convert it into the corresponding status type variable; Get the reservation status of each vacant room and convert it into the corresponding reservation type variable; Get the variable indicating whether there is a cleaning requirement in each guest room and convert it into a corresponding requirement Boolean variable; The position numerical feature, the state type variable, the predetermined type variable and the required Boolean variable are stored as preset structure data to form the state information.

3. The guest room cleaning sequence generation method according to claim 2, characterized in that: The training samples are formed based on the historical guest room status information and the historical cleaning priority data, and the preset initial model is trained to obtain a multilayer perceptron, including: Extract historical guest room status information and match it with corresponding historical cleaning priority data to obtain a training set; Performing standardization processing on the position numerical features corresponding to the position information; Performing one-hot encoding on the state type variables, reservation type variables, and demand Boolean variables corresponding to the check-in information, reservation information, and customer demand information; Based on a preset loss function, the initial model is trained using the training set, the weights are adjusted through a back-propagation algorithm, and the loss function is minimized to obtain a final multi-layer perceptron.

4. The guest room cleaning sequence generation method according to claim 3, characterized in that: The step of processing the state information of each guest room based on the multi-layer perceptron and outputting the guest room cleaning sequence comprises: Input the currently acquired guest room status information into the trained multi-layer perceptron; Calculating the cleaning priority level of the corresponding guest room in the multi-layer perceptron; Based on the cleaning priority levels, all rooms are sorted to generate a priority list of cleaning tasks; The cleaning priority level is set as a discrete category label or a continuous priority score.

5. The guest room cleaning sequence generation method according to claim 4, wherein: When the cleaning priority level is set to discrete category labels, the loss function is set to a cross entropy loss function; when the cleaning priority level is set to a continuous priority score, the loss function is set to a mean square error loss function.

6. The guest room cleaning sequence generation method according to claim 4, characterized in that: When the cleaning priority level is set as a discrete category label, if there are rooms with the same label, the rooms with the same label are clustered, and the priority order of the rooms with the same label is output based on the clustering result.

7. The guest room cleaning sequence generation method according to claim 1, characterized in that: The method further comprises: The guest room cleaning sequence is sent to a preset mobile terminal, and the mobile terminal is used to assist cleaning personnel in performing guest room cleaning operations based on the guest room cleaning sequence.

8. A guest room cleaning sequence generation system, characterized in that: The system comprises: A status acquisition module, at least used to acquire status information of each guest room, wherein the status information includes location information, check-in information, reservation information and customer demand information; A model training module, at least used to form a training sample based on historical guest room status information and historical cleaning priority data, and train a preset initial model to obtain a multi-layer perceptron; The sequence generation module is at least used to process the status information of each guest room based on the multi-layer perceptron and output the guest room cleaning sequence.

9. The guest room cleaning sequence generation system according to claim 8, characterized in that: The system further comprises: The mobile terminal is at least used to obtain the room cleaning sequence generated by the sequence generation module.

10. A readable storage medium, characterized in that: The readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the guest room cleaning sequence generating method according to any one of claims 1 to 7 is implemented.

Images