Genetic algorithm-based old-age care meal assistance platform distribution resource optimization method and system

Abstract

The invention relates to an old-age care meal-assisting platform distribution resource optimization method and system based on a genetic algorithm, and the method comprises the following steps: obtaining a real-time order data stream, marking delay data, and carrying out the real-time analysis to obtain a real-time analysis result, the delay data represents orders which cannot be arrived or processed according to an expected time window and related data; analyzing historical data based on an LSTM model, analyzing user preferences and predicting a future order trend; and based on the real-time analysis result, the user preference and the future order trend, calculating a preliminary delivery scheme by adopting linear programming, and optimizing the delivery scheme by utilizing a genetic algorithm. Compared with the prior art, the method has the advantages of timely response, high resource scheduling efficiency and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution resource optimization, and particularly to a method and system for optimizing distribution resources of a pension meal assistance platform based on a genetic algorithm. Background Art

[0002] With the intensification of the global population aging trend, the demand for pension services has grown rapidly. As an important part of the daily life of the elderly, meal assistance services face multiple challenges such as expanding service scope, diverse demands, and optimizing distribution efficiency. Traditional meal assistance services mainly rely on telephone reservations or offline meal ordering, making it difficult to meet the growing personalized needs of the elderly population.

[0003] Currently, the main modes of meal assistance services in the market include:

[0004] 1. Centralized meal assistance points: The elderly go to designated locations for meals, but there are limitations for the elderly with mobility difficulties.

[0005] 2. Offline delivery mode: Ordering and delivery are completed through the organization of community personnel, but the order response and resource scheduling efficiency are low.

[0006] 3. Simple online platform: An online meal assistance system based on fixed menus and simplified scheduling, but lacking intelligence and large-scale processing capabilities.

[0007] The above existing technologies have the following deficiencies:

[0008] 1. Delayed response: The traditional system lacks the ability to process real-time data streams, resulting in delayed order distribution and inability to meet peak demands.

[0009] 2. Inefficient resource scheduling: The distribution routes and resource allocation are based on fixed rules, lacking intelligent optimization means, resulting in resource waste.

[0010] 3. Poor user experience: It is unable to dynamically recommend meals according to user preferences and is difficult to meet personalized needs. Summary of the Invention

[0011] The purpose of the present invention is to overcome the defects of the above existing technologies and provide a method and system for optimizing distribution resources of a pension meal assistance platform based on a genetic algorithm, which combines real-time data stream processing, intelligent resource scheduling, and personalized recommendation technologies to solve the deficiencies of the existing modes and improve the efficiency of meal assistance services and user satisfaction.

[0012] The purpose of the present invention can be achieved through the following technical solutions:

[0013] A method for optimizing distribution resources of a pension meal assistance platform based on a genetic algorithm, the method comprising the following steps:

[0014] Obtain the real-time order data stream, mark the delayed data, and conduct real-time analysis to obtain the real-time analysis results. Among them, the delayed data represents orders and related data that fail to arrive or be processed within the expected time window;

[0015] Analyze historical data based on the LSTM model, predict future order trends, and analyze user preferences;

[0016] Based on the real-time analysis results, user preferences, and future order trends, use linear programming to calculate the preliminary delivery plan and optimize the delivery plan using the genetic algorithm.

[0017] The delayed data includes:

[0018] Late order data: User order data that is not uploaded to the elderly meal assistance platform within the specified time window;

[0019] Delay in delivery status update: Status data of deliverymen that is not updated in a timely manner;

[0020] External data delay: Environmental data that fails to arrive on time due to API delay or system failure;

[0021] Merchant data delay: Meal inventory or menu data of meal assistance points that is not updated on time.

[0022] The real-time analysis is to perform dynamic calculation and comprehensive analysis on the data stream within the current time period based on the sliding window algorithm, including order trend analysis, delivery efficiency analysis, resource utilization analysis, user behavior analysis, and anomaly analysis; among them,

[0023] The order trend analysis includes:

[0024] Order volume change analysis: Analyze the total number of orders within the current time period and its change trend compared with the previous time window;

[0025] Peak time prediction: Predict the future order peak time;

[0026] Regional order distribution: Analyze the proportion of orders in different regions and mark the key delivery regions;

[0027] The delivery efficiency analysis includes:

[0028] Average delivery time statistics: Statistically analyze the average delivery duration of orders within the current time period;

[0029] Delayed order proportion analysis: Mark the proportion of delayed orders and analyze the reasons;

[0030] The resource utilization analysis includes:

[0031] Analyze the load situation of deliverymen: Determine the current order quantity and load level of each deliveryman;

[0032] Analyze vehicle resource allocation: Count the utilization rate of distribution vehicles;

[0033] The user behavior statistics include:

[0034] User activity analysis: Count the number of users who place orders and their repeat order situations during the current time period;

[0035] Cancellation rate analysis: Count the proportion and reasons for users to cancel orders during the current time period;

[0036] The anomaly analysis includes:

[0037] Anomaly order statistics: Count the number and details of delayed orders or duplicate orders;

[0038] Distribution problem statistics: Count the number of problem orders and their reasons caused by abnormal paths or failed delivery staff status updates.

[0039] The historical data mentioned above includes:

[0040] User order data: Including the number of orders, order placement time, and order amount;

[0041] User preference data: Including dish selection, special dietary requirements, and taste preferences;

[0042] User usage habits: Including order placement frequency, active periods, and platform usage duration;

[0043] User basic information: Including user age, health status, and geographical location.

[0044] The specific steps for analyzing user preferences are as follows:

[0045] Extract user behavior features through data preprocessing and feature engineering. The user behavior features include order activity, user preference features, consumption habits, and time features. Among them, the order activity is determined based on the number of user orders and the most recent order placement time. The user preference features include the most frequently ordered dishes and specific dietary requirements. The consumption habits include the mean and distribution of order amounts. The time features include the peak order placement periods of users;

[0046] Normalize the extracted user behavior features;

[0047] Use the K-Means algorithm for clustering user behavior features:

[0048] Set the number of clustering centers;

[0049] Input the user behavior characteristics into the K-Means algorithm, calculate the Euclidean distance between each user and each cluster center, and assign the user to the nearest cluster center according to the distance.

[0050] Continuously iterate and update the positions of the cluster centers until the clustering result converges or reaches the maximum number of iterations.

[0051] Evaluate the quality of the clustering result, determine whether it is necessary to adjust the number of cluster centers or the user behavior characteristic data. If adjustment is needed, re-perform the user behavior characteristic clustering after adjustment; if not, output the clustering result.

[0052] The analysis result of the user preferences is to divide users into multiple types of groups, including health-needs users, mobility-limited users, social-needs users, economy-saving users, and occasional users. Among them, the characteristics of the health-needs users are that they have selected low-salt, low-fat, and low-sugar meals for ordering multiple times, or the order remarks include special dietary requirements; the characteristics of the mobility-limited users are that the order places are concentrated at a fixed home address, and they frequently choose the delivery service, and the order remarks include the need to deliver meals to home; the characteristics of the social-needs users are that they frequently choose to dine at the meal assistance points, and the order remarks include dining with others or having a dinner with fixed people; the characteristics of the economy-saving users are that the average order amount is lower than the per capita level or the order remarks include asking for preferential information / price reduction requests; the characteristics of the occasional users are that the order frequency is lower than the average level.

[0053] The specific steps for calculating the preliminary delivery plan using linear programming are as follows:

[0054] Order and delivery resource matching:

[0055] Analyze the order quantity and meal demand in each area, and preliminarily allocate the available delivery staff, vehicles, and delivery time periods.

[0056] Calculate the task volume of each delivery staff and adjust it according to the set upper and lower limits of the delivery staff load.

[0057] Delivery route planning:

[0058] Generate a preliminary delivery route plan based on the geographical location of the meal assistance points, the order delivery addresses, and the road condition data.

[0059] According to the area priority principle, group the orders within a preset distance range and assign the same delivery staff to be responsible.

[0060] Time window allocation:

[0061] Allocate the corresponding delivery time window for each order according to the user preferences to ensure that high-priority orders are processed first.

[0062] Preliminarily determine whether the delivery task can be completed within the specified delivery time window. If it cannot be satisfied, mark it as needing optimization.

[0063] The specific method of optimizing the delivery plan using the genetic algorithm is as follows: Based on the preliminary delivery plan, by simulating the process of natural selection and genetic evolution, optimize the resource allocation, delivery route, and time window allocation. Among them, for resource allocation optimization, on the basis of the preliminary plan, taking the balance degree of delivery tasks and resource utilization rate as the fitness function, use the genetic algorithm to find the delivery plan with the shortest path length and the lowest time cost; for resource allocation optimization, on the basis of the resource allocation of delivery personnel, vehicles, and meal assistance points, taking the total path length, time cost, or fuel consumption as the fitness function, use the genetic algorithm to optimize the resource utilization rate; for time window matching optimization, according to the priority of orders, taking the balance of the on-time delivery rate of orders and the total delivery time as the fitness function, use the genetic algorithm to dynamically adjust the time sequence of delivery tasks.

[0064] The method further includes:

[0065] According to the predicted future order trend, allocate the types and quantities of meals to be prepared at the meal assistance points;

[0066] For the allocation of the stock quantity of meals at the meal assistance points, taking the inventory utilization rate and the matching degree of the stock quantity and demand as the fitness function, use the genetic algorithm to optimize the meal preparation efficiency and inventory.

[0067] A delivery resource optimization system for an elderly meal assistance platform based on the genetic algorithm, used to implement the above-mentioned method. The system includes:

[0068] Real-time data acquisition and analysis module: Acquire the real-time order data stream and mark the delayed data, and perform real-time analysis to obtain the real-time analysis result. Among them, the delayed data represents the orders and related data that fail to arrive or be processed within the expected time window;

[0069] Historical data analysis module: Analyze historical data based on the LSTM model, analyze user preferences, and predict future order trends;

[0070] Delivery plan planning module: Based on the real-time analysis result, user preferences, and future order trends, use linear programming to calculate the preliminary delivery plan, and use the genetic algorithm to optimize the delivery plan.

[0071] Compared with the prior art, the present invention has the following beneficial effects:

[0072] (1) Fast response: The present invention considers the influence of delayed data in the acquired real-time data, performs real-time data analysis, can reduce order delays, and improve the response speed.

[0073] (2) Intelligence: The present invention conducts predictive analysis based on historical data, predicts future order demands and user preferences, and optimizes the scheduling of distribution resources based on user preferences, capable of meeting personalized needs.

[0074] (3) Efficiency: The present invention uses a genetic algorithm to optimize resource allocation, distribution routes, and time window allocation, reducing the distribution time cost, improving the distribution efficiency, and ensuring the rationality of the task arrangements for each delivery person and maximizing the overall distribution task efficiency while meeting high-priority orders. Description of the Drawings

[0075] Figure 1 It is a flowchart of the method of the present invention. Detailed Embodiments

[0076] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0077] Unless otherwise defined, the technical terms or scientific terms involved in this application shall have the ordinary meaning understood by those of ordinary skill in the technical field to which this application belongs. The "one", "a", "an", "the" and other similar words involved in this application do not represent a quantity limitation and can represent a singular or plural number. The terms "including", "comprising", "having" and any variations thereof involved in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product or device including a series of steps or modules (units) is not limited to the listed steps or units, but may further include unlisted steps or units, or may further include other steps or units inherent to these processes, methods, products or devices. The "connection", "coupling" and other similar words involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The "multiple" involved in this application refers to two or more. The "and / or" describes the association relationship of associated objects and indicates that three relationships may exist. For example, "A and / or B" may represent: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally represents an "or" relationship between the associated objects before and after. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific order for the objects.

[0078] This embodiment provides an optimization method for distribution resources of a pension meal assistance platform based on a genetic algorithm, asFigure 1 As shown, the method includes the following steps:

[0079] S1. Obtain the real-time order data stream and mark the delayed data, and perform real-time analysis to obtain the real-time analysis result.

[0080] Delayed data refers to orders and related data that fail to arrive or be processed within the expected time window during data stream processing. Specifically, it includes the following types:

[0081] 1) Late order data: User order data that is not uploaded to the elderly meal assistance platform within the specified time window, such as due to unstable network or delayed client submission.

[0082] Example: The user places an order at 10:00 am, but due to network problems, the data is not uploaded until 10:15.

[0083] 2) Delay in delivery status update: Status data of deliverymen that is not updated in a timely manner, which may lead to inaccurate real-time scheduling.

[0084] Example: The deliveryman fails to mark the delivery in a timely manner after completing the order, affecting the system's analysis of delivery efficiency.

[0085] 3) External data delay: Environmental data (such as real-time weather and road condition data) that fails to arrive on time due to API delay or system failure.

[0086] Example: The road condition information is not updated in a timely manner, resulting in the failure to optimize the route plan in real time.

[0087] 4) Merchant data delay: Meal inventory or menu data of meal assistance points that is not updated on time, affecting the accuracy of order allocation.

[0088] Example: The merchant updates the menu at 8:00 am, but the system does not obtain the data until 8:30.

[0089] Real-time analysis is to perform dynamic calculation and comprehensive analysis on the data stream within the current time period based on the sliding window algorithm, including:

[0090] 1) Order trend analysis

[0091] Analysis of order volume change: Analyze the total number of orders within the current time period and its change trend (increase, flat, or decrease) compared with the previous time window;

[0092] Peak time prediction: Predict the future order peak time to help schedule resources in advance;

[0093] Regional order distribution: Analyze the proportion of orders in different regions and mark the key delivery regions.

[0094] 2) Delivery efficiency analysis

[0095] Average delivery time statistics: Statistically analyze the average delivery duration of orders within the current time period;

[0096] Analysis of the proportion of delayed orders: Mark the proportion of delayed orders and analyze the reasons (such as traffic congestion, overloaded delivery staff, etc.).

[0097] 3) Analysis of resource utilization

[0098] Analysis of delivery staff load: Determine the current number of orders and the load level of each delivery staff;

[0099] Analysis of vehicle resource allocation: Statistically analyze the utilization rate of delivery vehicles.

[0100] 4) User behavior statistics

[0101] Analysis of user activity: Statistically analyze the number of users who place orders and their repeat order placement within the current time period;

[0102] Analysis of cancellation rate: Statistically analyze the proportion of orders cancelled by users and the reasons (such as timeout, out of stock, etc.) within the current time period.

[0103] 5) Abnormality analysis

[0104] Statistics of abnormal orders: Statistically analyze the number and details of delayed orders or duplicate orders;

[0105] Statistics of delivery problems: Statistically analyze the number of problem orders caused by abnormal routes and failed updates of delivery staff status and their reasons.

[0106] S2. Analyze historical data based on the LSTM model, predict future order trends, and analyze user preferences.

[0107] In this embodiment, the historical data includes:

[0108] User order data: including the number of orders, order placement time, order amount, etc.;

[0109] User preference data: including dish selection, special dietary requirements (such as low salt, low sugar), and taste preferences;

[0110] User usage habits: including order placement frequency, active periods (morning, noon, evening), and platform usage duration;

[0111] User basic information: including user age, health status, and geographical location, etc.

[0112] The LSTM network is a network model used for time series prediction. In this embodiment, the part of analyzing historical data based on the LSTM model and predicting future order trends will not be elaborated further.

[0113] Specifically, analyzing user preferences includes the following steps:

[0114] S21. Extract user behavior features through data preprocessing and feature engineering.

[0115] User behavior features include order activity, user preference features, consumption habits, and time features. Among them, order activity is determined based on the number of user orders and the time of the most recent order. User preference features include the most frequently ordered dishes and specific dietary requirements (such as low-fat meals, vegetarian meals). Consumption habits include the mean and distribution of order amounts. Time features include the peak order periods of users (such as lunch or dinner).

[0116] S22. Normalize the extracted user behavior features to convert data with different units and ranges into a unified range (such as 0 to 1) to avoid the excessive influence of certain features on the clustering results.

[0117] S23. Use the K-Means algorithm for clustering user behavior features:

[0118] S231. Set the number of cluster centers, which can be determined by the elbow method in this embodiment.

[0119] S232. Input the user behavior features into the K-Means algorithm, calculate the Euclidean distance between each user and each cluster center, and assign it to the nearest cluster center according to the distance.

[0120] S233. Continuously iterate and update the positions of the cluster centers until the clustering results converge or reach the maximum number of iterations.

[0121] S234. Use metrics such as the Silhouette Score to evaluate the quality of the clustering results, and judge whether it is necessary to adjust the number of cluster centers or the user behavior feature data. If adjustment is needed, re-cluster the user behavior features after adjustment. If not, output the clustering results.

[0122] The analysis result of user preferences is to divide users into multiple types of groups, including:

[0123] Type A: Users with health needs

[0124] Features: Frequently choose low-salt, low-fat, and low-sugar meals, or note special dietary requirements (such as diabetes, hypertension, cardiovascular diseases, etc.).

[0125] Needs:

[0126] Pay attention to diet health and preferentially choose diets that match their own health conditions.

[0127] Tend to customized nutrition packages or healthy recommended dishes.

[0128] Service optimization suggestions:

[0129] Launch a healthy dining zone and regularly update the list of low-salt and low-sugar dishes.

[0130] Provide nutritional consultation services or personalized healthy meal plans.

[0131] Category B: Users with limited mobility

[0132] Characteristics: The order placement locations are concentrated at fixed home addresses, frequently choose takeout services, and the remarks often mention "need to deliver food to home".

[0133] Needs:

[0134] Hope that the meals can be safely and punctually delivered to home to reduce the burden of going out.

[0135] Pay attention to the delivery timeliness and service attitude.

[0136] Service optimization suggestions:

[0137] Improve the coverage of takeout services, especially in remote areas.

[0138] Optimize the delivery process and provide "contactless delivery" or "delivery to the floor" services.

[0139] Category C: Users with social needs

[0140] Characteristics: Habitually choose to dine at the meal assistance points, often remark that they are dining with others or having a dinner with fixed people.

[0141] Needs:

[0142] Meal assistance is not only about dining but also a part of social activities. Hope to have a comfortable dining environment and rich social activities.

[0143] Service optimization suggestions:

[0144] Organize theme activities (such as health lectures, interest groups) at the meal assistance points to attract users.

[0145] Optimize the dining environment (such as setting up tables for the elderly, creating a barrier-free environment).

[0146] Category D: Economical users

[0147] Characteristics: Tend to choose economical packages, pay attention to meal assistance subsidies or preferential activities, and often remark "cheaper" or ask about subsidy policies.

[0148] Needs:

[0149] Price-sensitive and more concerned about preferential policies and subsidies.

[0150] Service optimization suggestions:

[0151] Launch cost-effective packages or offer discounts during special periods (such as off-peak dining).

[0152] Clearly publicize the meal subsidy policy to increase user participation.

[0153] Category E: Occasional users

[0154] Characteristics: Lower order frequency, mainly concentrated on holidays or specific periods (such as when family members go out).

[0155] Needs:

[0156] Meal assistance is a supplementary service and hopes to provide efficient services during specific periods.

[0157] Service optimization suggestions:

[0158] Push holiday discount packages or launch a flexible ad-hoc meal ordering service.

[0159] S3. Based on the real-time analysis results, user preferences and future order trends, use linear programming to calculate a preliminary delivery plan and optimize the delivery plan using a genetic algorithm.

[0160] S31. Calculate a preliminary delivery plan using linear programming

[0161] S311. Match orders with delivery resources:

[0162] Analyze the order quantity and meal requirements in each area, and initially allocate available delivery staff, vehicles and delivery time slots;

[0163] Calculate the workload of each delivery staff and adjust it according to the set upper and lower limits of the delivery staff's load to ensure balanced allocation and avoid overloading or idling of individual resources.

[0164] S312. Delivery route planning:

[0165] Based on the geographical location of the meal assistance points, the order delivery addresses and road condition data, generate a preliminary delivery route plan (not optimal but basically reasonable);

[0166] According to the regional priority principle, group orders within a preset distance range and assign the same delivery staff to be responsible.

[0167] S313. Time window allocation:

[0168] Allocate corresponding delivery time windows for each order according to user preferences to ensure that high-priority orders (such as during peak periods or urgent needs) are processed first;

[0169] Preliminarily judge whether the delivery task can be completed within the specified delivery time window. If it cannot be satisfied, mark it as needing optimization.

[0170] S314, Allocation of Meal-Assistance Ordering Requirements:

[0171] Based on the predicted order demands, allocate the types and quantities of meals to be prepared at the meal-assistance points to ensure sufficient inventory without excessive waste.

[0172] S315, Evaluation of Resource Utilization Rate:

[0173] Calculate the resource utilization rate based on the current resource availability (such as delivery staff, vehicles, capacity of meal-assistance points, etc.) and preliminarily identify possible resource bottleneck areas.

[0174] For example, the meal-assistance service demands in a certain area are as follows:

[0175] (1) Order demand data:

[0176] Area A: 100 orders (30 delivery addresses are close, 50 are of medium distance, 20 are of long distance)

[0177] Area B: 60 orders (50 are close, 10 are of medium distance)

[0178] Area C: 40 orders (all are close)

[0179] (2) Delivery staff and vehicle resources:

[0180] Available delivery staff: 6 people

[0181] Delivery vehicles: 3 vehicles

[0182] (3) Meal-assistance point inventory data:

[0183] Meal-assistance point X: The inventory can provide 120 meals.

[0184] Meal-assistance point Y: The inventory can provide 80 meals.

[0185] The preliminary allocation plan generated is as follows:

[0186] (1) Matching of orders and delivery resources:

[0187] Allocation of delivery staff:

[0188] Area A: Allocate 3 delivery staff (each delivers an average of 33 - 34 orders)

[0189] Area B: Allocate 2 delivery staff (each delivers 30 orders)

[0190] Area C: Allocate 1 delivery staff (delivers 40 orders)

[0191] Allocation of vehicles:

[0192] Area A: 2 delivery vehicles (responsible for medium and long-distance order deliveries)

[0193] Area B: 1 delivery vehicle (responsible for medium-distance order deliveries)

[0194] Area C: No vehicle (short-distance orders can be completed on foot)

[0195] (2) Delivery route planning:

[0196] Area A:

[0197] The first group of routes: Responsible for 30 short-distance orders, completed by deliverymen on foot.

[0198] The second group of routes: Responsible for 50 medium-distance orders, with the vehicle connecting order addresses by the shortest path.

[0199] The third group of routes: Responsible for 20 long-distance orders, with the vehicle planning a better route according to the road conditions.

[0200] Area B:

[0201] The first group of routes: 50 short-distance orders, completed on foot.

[0202] The second group of routes: 10 medium-distance orders, delivered by vehicle.

[0203] Area C:

[0204] All 40 short-distance orders, completed by deliverymen on foot.

[0205] (3) Time window allocation:

[0206] High-priority orders (such as meals for the elderly that need to be delivered before 12:00): Medium and long-distance orders in Area A are preferentially assigned to the second and third groups of routes.

[0207] Regular orders: Time windows are divided according to the estimated delivery time (such as 10 - 12 am, 1 - 3 pm, etc.).

[0208] (4) Meal delivery order allocation:

[0209] Meal delivery point X: Meets the order requirements of Area A, and 100 meals need to be prepared.

[0210] Meal delivery point Y: Meets the requirements of Area B and Area C, and 100 meals need to be prepared.

[0211] (5) Resource utilization rate assessment:

[0212] Deliveryman utilization rate: The daily task volume of each deliveryman is between 30 and 40 orders, and the task distribution is relatively balanced.

[0213] Vehicle utilization rate: The utilization rate of 3 delivery vehicles is 90%, meeting the current demand.

[0214] Meal assistance point inventory: The inventory at meal assistance points X and Y can basically meet the order demand, but it needs to be replenished in a timely manner before the service peak.

[0215] S32. Optimize the delivery plan using the genetic algorithm

[0216] The genetic algorithm is mainly used to deeply optimize resource allocation, delivery routes, and overall scheduling on the basis of the preliminary allocation plan by simulating the processes of natural selection and genetic evolution, further improving system efficiency, reducing costs, and ensuring service quality.

[0217] The goal of genetic algorithm optimization is to find the global optimal solution for resource allocation and delivery routes, including but not limited to the following aspects:

[0218] a. Delivery route optimization: On the basis of the preliminary planning, further search for a delivery plan with the shortest route length and the lowest time cost.

[0219] Delivery route optimization is the core application content of the genetic algorithm. Its goal is to find the optimal delivery routes for delivery personnel and vehicles to minimize the total delivery time, distance, or cost.

[0220] Optimization problems: i. Route planning for each delivery personnel and vehicle to ensure that all orders are effectively covered. ii. The total route distance is the shortest, considering traffic congestion and actual road conditions.

[0221] Fitness function: Evaluate using the total route length, time cost, or fuel consumption as the fitness function.

[0222] Optimization effect: Through the genetic algorithm, find a delivery route that is better than the preliminary route planning, reduce the time cost, and improve the efficiency.

[0223] Example: The route initially assigned to delivery personnel A covers 20 orders, and the total route length is 50 kilometers. After optimization by the genetic algorithm, the total route length is shortened to 40 kilometers.

[0224] In this embodiment, after optimization by the genetic algorithm, the average route length is reduced by 10% - 20%, and the delivery time is shortened by 15% - 25%.

[0225] b. Resource allocation optimization: On the basis of the resource allocation of delivery personnel, vehicles, and meal assistance points, optimize the resource utilization rate to avoid resource waste or overload.

[0226] Optimization problems: i. Balance the task loads of delivery personnel and vehicles to avoid overload or idleness of some resources. ii. Reasonably allocate the inventory at meal assistance points to avoid resource waste.

[0227] Fitness function: The balance degree of distribution tasks and inventory utilization rate are used as the fitness function.

[0228] Optimization effect: Ensure that the task volume of each delivery person or vehicle is moderate, and at the same time reduce the inventory waste at the meal assistance points.

[0229] In this embodiment, after being optimized by the genetic algorithm, the task loads of each delivery person and vehicle are more balanced, reducing the risk of individual overload.

[0230] c. Time window matching optimization: Dynamically adjust the time sequence of distribution tasks according to the priority of orders (such as the meal delivery time specified by the elderly).

[0231] Optimization problem: How to maximize the overall distribution task efficiency while meeting high-priority orders.

[0232] Fitness function: The balance between the order on-time delivery rate and the total distribution time is used as the fitness function.

[0233] Optimization effect: Improve the completion rate of high-priority orders (such as delivery at the specified time), and at the same time reduce the delay rate of ordinary orders.

[0234] In this embodiment, after being optimized by the genetic algorithm, the on-time delivery rate of high-priority orders is increased to over 95%.

[0235] d. Stock preparation allocation optimization: Optimize the stock preparation quantity of meal items at the meal assistance points to avoid waste or insufficient inventory.

[0236] Optimization problem: How to dynamically allocate the stock preparation quantity at the meal assistance points according to order demands to avoid insufficient or excessive meal items.

[0237] Fitness function: The inventory utilization rate and the matching degree between the stock preparation quantity and the demand are used as the fitness function.

[0238] Optimization effect: Dynamically adjust the inventory at the meal assistance points to achieve precise matching between the stock preparation quantity and the actual demand.

[0239] In this embodiment, after being optimized by the genetic algorithm, the stock preparation waste rate at the meal assistance points is reduced by 10% - 15%.

[0240] S4. Automated deployment: Use Kubernetes to expand container instances and automatically adjust computing resources according to the load.

[0241] S5. Data visualization: Use ECharts or Tableau to generate analysis charts such as order completion rate and distribution efficiency.

[0242] This embodiment also provides a distribution resource optimization system for a pension meal assistance platform based on the genetic algorithm, used to implement the above-mentioned method. The system includes:

[0243] ① Real-time data acquisition and analysis module: It acquires the real-time order data stream, marks the delayed data, and conducts real-time analysis to obtain real-time analysis results. In this embodiment, Apache Flink is used for real-time order processing to ensure the efficient access and analysis of the data stream.

[0244] ② Historical data analysis module: It analyzes historical data based on the LSTM model, analyzes user preferences, and predicts future order trends.

[0245] ③ Delivery plan planning module: Based on the real-time analysis results, user preferences, and future order trends, it calculates the preliminary delivery plan using linear programming and optimizes the delivery plan using the genetic algorithm.

[0246] ④ Automated deployment module: It realizes elastic expansion and load balancing through Kubernetes to ensure the stable operation of the system during peak periods.

[0247] ⑤ Visualization and interaction module: It provides an intuitive user interface and supports dynamic order monitoring and personalized recommendations.

[0248] In this embodiment, a reliable and efficient cloud storage and management system is constructed for data storage and management. Among them, data storage: Based on a distributed database (such as AWS S3 or Azure Blob), it stores user, merchant, and logistics data. Data management: Through a cloud service platform (such as AWS RDS or Azure SQL), it supports structured queries and data indexing. Data synchronization: It uses a distributed cache (such as Redis) to achieve fast data synchronization.

[0249] The visualization and interaction module provides the following functions:

[0250] Data visualization: It generates data charts such as order volume, delivery efficiency, and user satisfaction through ECharts or Tableau.

[0251] User feedback: It establishes a user-side evaluation mechanism to collect service satisfaction and improvement suggestions.

[0252] Management decision support: It provides data reports and optimization suggestions (such as increasing delivery resources during specific periods).

[0253] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the described modules can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0254] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for optimizing distribution resources of a senior care meal assistance platform based on genetic algorithm, characterized in that: The method comprises the following steps: Acquire the real-time order data stream and mark the delayed data, and perform real-time analysis to obtain real-time analysis results, wherein the delayed data represents orders and related data that failed to arrive or be processed according to the expected time window; Analyze historical data based on the LSTM model, predict future order trends, and analyze user preferences; Based on real-time analysis results, user preferences and future order trends, linear programming is used to calculate the preliminary delivery plan, and genetic algorithm is used to optimize the delivery plan.

2. According to claim 1, a method for optimizing distribution resources of a meal assistance platform for the elderly based on a genetic algorithm is characterized in that: The delay data includes: Late order data: user order data that was not uploaded to the elderly care and dining assistance platform within the specified time window; Delay in delivery status update: The delivery person’s status data is not updated in time; External data delay: environmental data that fails to arrive on time due to API delays or system failures; Merchant data delay: The food inventory or menu data of the dining point is not updated on time.

3. According to claim 1, a method for optimizing distribution resources of a meal assistance platform for the elderly based on a genetic algorithm is characterized in that: The real-time analysis is to dynamically calculate and comprehensively analyze the data flow in the current time period based on the sliding window algorithm, including order trend analysis, delivery efficiency analysis, resource utilization analysis, user behavior analysis and anomaly analysis; The order trend analysis includes: Order volume change analysis: Analyze the total number of orders in the current time period and its change trend compared with the previous time window; Peak time prediction: predict future order peak times; Regional order distribution: Analyze the order proportions in different regions and mark key delivery areas; The distribution efficiency analysis includes: Average delivery time statistics: Statistics on the average delivery time of orders in the current time period; Analysis of delayed order ratio: mark the ratio of delayed orders and analyze the reasons; The resource utilization analysis includes: Analyze the delivery staff’s load: determine the current number of orders and load level of each delivery staff; Analyze vehicle resource allocation: Count the utilization rate of distribution vehicles; The user behavior statistics include: User activity analysis: count the number of users who have placed orders in the current time period and their repeated orders; Cancellation rate analysis: Statistics on the proportion and reasons of users canceling orders in the current time period; The abnormality analysis includes: Abnormal order statistics: count the number and details of delayed orders or repeated orders; Delivery problem statistics: Statistics on the number of problem orders and their causes caused by abnormal routes and delivery personnel status update failures.

4. According to the method for optimizing distribution resources of the elderly care meal assistance platform based on genetic algorithm in claim 1, it is characterized in that: The historical data include: User order data: including order quantity, order time, and order amount; User preference data: including menu selection, special dietary needs and taste preferences; User usage habits: including order frequency, active time periods, and platform usage duration; User basic information: including user age, health status and geographic location.

5. According to the method for optimizing distribution resources of the elderly care meal assistance platform based on genetic algorithm in claim 1, it is characterized in that: The analyzing of user preferences specifically includes the following steps: Extracting user behavior features through data preprocessing and feature engineering, wherein the user behavior features include order activity, user preference features, consumption habits and time features, wherein the order activity is determined based on the number of orders placed by the user and the time of the most recent order, the user preference features include the most frequently ordered dishes and specific dietary requirements, the consumption habits include the mean and distribution of the order amount, and the time features include the peak ordering period of the user; Normalize the extracted user behavior features; Use K-Means algorithm to cluster user behavior features: Set the number of cluster centers; Input user behavior features into the K-Means algorithm, calculate the Euclidean distance between each user and each cluster center, and assign them to the nearest cluster center based on the distance; The location of the cluster center is continuously updated iteratively until the clustering result converges or the maximum number of iterations is reached; Evaluate the quality of the clustering results and determine whether it is necessary to adjust the number of cluster centers or user behavior feature data. If adjustment is required, re-cluster the user behavior features after the adjustment. If not, output the clustering results.

6. The method for optimizing distribution resources of a meal assistance platform for the elderly based on a genetic algorithm according to claim 1, characterized in that: The analysis result of the user preferences is to divide users into multiple types of groups, including users with health needs, users with limited mobility, users with social needs, economical users and occasional users, among which the users with health needs are characterized by repeatedly choosing low-salt, low-fat, and low-sugar food orders, or including special dietary needs in the order notes; the users with limited mobility are characterized by ordering locations concentrated at fixed home addresses, and frequently choosing delivery services, with the order notes including the need for home delivery; the users with social needs are characterized by frequently choosing to dine at assisted dining points, with the order notes including traveling with other people or dining with fixed people; the economical users are characterized by the average order amount being lower than the per capita level or the order notes including inquiries about discount information / price reduction requests; the occasional user is characterized by an order frequency lower than the average level.

7. According to claim 1, a method for optimizing distribution resources of a senior care meal assistance platform based on a genetic algorithm is characterized in that: The method of calculating the preliminary distribution plan by linear programming specifically includes the following steps: Matching orders with delivery resources: Analyze the order quantity and food demand in each area, and preliminarily allocate available delivery personnel, vehicles and delivery time periods; Calculate the workload of each delivery person and adjust it according to the set upper and lower limits of the delivery person's load; Delivery route planning: Generate a preliminary delivery route plan based on the geographical location of the meal delivery point, the order delivery address and road condition data; According to the regional priority principle, orders within a preset distance are grouped together and assigned to the same delivery person; Time window allocation: Assign a corresponding delivery time window to each order based on user preferences to ensure that high-priority orders are processed first; Preliminary judgment is made on whether the delivery task can be completed within the specified delivery time window. If it cannot be met, it is marked as requiring optimization.

8. The method for optimizing distribution resources of a meal assistance platform for the elderly based on a genetic algorithm according to claim 1, characterized in that: The use of genetic algorithms to optimize the distribution plan is specifically as follows: on the basis of the preliminary distribution plan, by simulating the process of natural selection and genetic evolution, the resource allocation, distribution path and time window allocation are optimized, wherein the resource allocation optimization is based on the preliminary plan, with the balance of distribution tasks and resource utilization as the fitness function, and the genetic algorithm is used to find the distribution plan with the shortest path length and the lowest time cost; the resource allocation optimization is based on the allocation of distribution personnel, vehicles and meal assistance points, with the total path length, time cost or fuel consumption as the fitness function, and the genetic algorithm is used to optimize the resource utilization; the time window matching optimization is based on the order priority, with the balance between the on-time delivery rate of the order and the total delivery time as the fitness function, and the genetic algorithm is used to dynamically adjust the time sequence of the distribution tasks.

9. The method for optimizing distribution resources of a meal assistance platform for the elderly based on a genetic algorithm according to claim 1, characterized in that: The method further comprises: According to the predicted future order trends, the types and quantities of meals that need to be prepared are allocated to the catering stations; For the allocation of stock for meal ordering, the efficiency and inventory of meal preparation are optimized using genetic algorithm, taking inventory utilization rate and the matching degree between stock quantity and demand as fitness function.

10. A distribution resource optimization system for elderly care meal assistance platform based on genetic algorithm, characterized in that: For implementing the method according to any one of claims 1 to 9, the system comprises: Real-time data acquisition and analysis module: acquires real-time order data stream and marks delayed data, performs real-time analysis to obtain real-time analysis results, wherein the delayed data represents orders and related data that failed to arrive or be processed according to the expected time window; Historical data analysis module: Analyze historical data based on the LSTM model, analyze user preferences and predict future order trends; Delivery plan planning module: Based on real-time analysis results, user preferences and future order trends, linear programming is used to calculate the preliminary delivery plan, and genetic algorithms are used to optimize the delivery plan.

Images