Big data driven hotel dynamic pricing and revenue optimization intelligent system

Abstract

This big data-driven intelligent system for dynamic hotel pricing and revenue optimization, encompassing the field of business management technology, extracts event scores, weather scores, and traffic scores. These scores are then ranked to obtain a first demand index, a second demand index, and a third demand index. A dynamic demand index is obtained by combining these scores with preset fusion weights. A pricing adjustment coefficient is output through a preset pricing mapping model. An initial price is generated by combining this with the base price of the target room type. The initial price is then adjusted based on competitor pricing and a preset sensitivity coefficient to obtain the final price. An order volatility index is obtained after implementing the final price, and the preset fusion weights or preset sensitivity coefficient are updated accordingly. The system can integrate multi-dimensional heterogeneous data in real time to generate accurate demand forecasting indices, thereby dynamically optimizing pricing strategies and improving hotel revenue potential.

Description

Technical Field

[0001] This application relates to the field of business management technology, specifically a big data-driven intelligent system for dynamic hotel pricing and revenue optimization. Background Technology

[0002] Traditional hotel revenue management relies mainly on historical occupancy data and explicit pricing information from competitors for decision-making. This approach has certain limitations when dealing with regular and predictable demand patterns, especially when facing sudden and short-term fluctuations in local demand, which often leads to pricing strategies failing to respond in a timely manner.

[0003] For example, hotels may miss revenue opportunities during peak demand periods due to excessively low pricing, while facing the risk of vacant rooms during off-peak demand periods due to excessively high pricing. Therefore, hotel managers find it difficult to balance demand capture and competitiveness in a rapidly changing market environment, resulting in the overall revenue potential not being fully realized. Existing technologies urgently need to be improved to address these issues. Summary of the Invention

[0004] The purpose of this application is to provide a big data-driven intelligent system for dynamic pricing and revenue optimization of hotels, which can integrate multi-dimensional heterogeneous data in real time to generate accurate demand forecasting indices, thereby dynamically optimizing pricing strategies and improving the hotel's revenue potential.

[0005] The objective of this application can be achieved through the following technical solution: Firstly, a big data-driven intelligent system for dynamic hotel pricing and revenue optimization, comprising the following modules:

[0006] The data acquisition module is used to synchronously acquire urban event data, regional meteorological data, and hotel traffic data within a preset monitoring period;

[0007] The data evaluation module is used to extract event scores and weather scores based on the city event data and regional meteorological data, respectively, and to extract a traffic score based on the weather scores combined with hotel traffic data.

[0008] The data fusion module is used to sort the various scores to obtain the first demand index, the second demand index, and the third demand index, and to obtain the dynamic demand index by combining the preset fusion weights of the corresponding order.

[0009] The pricing generation module is used to input the dynamic demand index and the first demand index into a preset pricing mapping model to output a pricing adjustment coefficient, and generate a corresponding preliminary price by combining the basic price of the target room type.

[0010] The pricing correction module is used to obtain the competitor pricing of the target room type within a preset monitoring period, and to correct the preliminary pricing by combining it with a preset sensitivity coefficient corresponding to the score of the first demand index to obtain its final pricing.

[0011] The feedback optimization module is used to obtain the order volatility index of the target room type within a preset time window after the final pricing is executed, and to update the preset fusion weight or preset sensitivity coefficient according to the index until the order volatility index that meets the preset output conditions is obtained.

[0012] Secondly, the big data-driven intelligent method for dynamic hotel pricing and revenue optimization includes the following steps:

[0013] Simultaneously acquire urban event data, regional meteorological data, and hotel traffic data within a preset monitoring period;

[0014] Based on the city event data and regional meteorological data, event scores and meteorological scores are extracted respectively. Based on the meteorological scores and hotel traffic data, a traffic score is extracted.

[0015] The scores are ranked to obtain the first demand index, the second demand index, and the third demand index. The dynamic demand index is obtained by combining the preset fusion weights of the corresponding order.

[0016] The dynamic demand index and the first demand index are input into a preset pricing mapping model to output a pricing adjustment coefficient, which is then combined with the base price of the target room type to generate the corresponding preliminary price.

[0017] Obtain the competitor pricing of the target room type within a preset monitoring period, and adjust the preliminary pricing by combining it with the preset sensitivity coefficient of the score corresponding to the first demand index to obtain its final pricing;

[0018] Obtain the order volatility index of the target room type within a preset time window after the final pricing is executed, and update the preset fusion weight or preset sensitivity coefficient according to it until the order volatility index that meets the preset output conditions is obtained.

[0019] Thirdly, a computer storage medium stores computer-executable instructions, which, when executed, implement the big data-driven intelligent system for dynamic hotel pricing and revenue optimization described in the first aspect.

[0020] Compared with the prior art, the beneficial effects of this application are:

[0021] As can be seen from the above, this application generates a dynamic demand index and optimizes pricing by integrating urban events, regional weather, and hotel traffic data. It can respond to changes in market demand in real time and has the advantage of improving the potential for hotel revenue. It realizes the transformation from historical data-driven to real-time environmental perception-driven, and can proactively capture and quantify unconventional demand opportunities, making pricing decisions more accurate and agile, and providing hotel managers with an intelligent way to optimize revenue. Attached Figure Description

[0022] Figure 1 A schematic diagram of the modules of the big data-driven intelligent system for dynamic hotel pricing and revenue optimization in this application;

[0023] Figure 2 This diagram illustrates the steps of the big data-driven intelligent method for dynamic hotel pricing and revenue optimization proposed in this application. Detailed Implementation

[0024] The technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components described and shown in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but only to illustrate selected embodiments of this application.

[0025] Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this application. It should be noted that similar reference numerals and letters in the following figures indicate similar items. Therefore, once an item has been defined in one figure, it does not need to be further defined and explained in subsequent figures. The terms "first", "second", etc. are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance.

[0026] Traditional hotel revenue management relies primarily on historical occupancy data and explicit pricing information from competitors for decision-making, lacking the ability to detect sudden, short-term fluctuations in localized demand in real time. Such systems fail to effectively integrate multi-source, heterogeneous implicit demand signals, causing predictive models to be unable to accurately quantify changes in actual market demand, resulting in a disconnect between pricing strategies and actual demand. The root of the problem lies in the single data source and lagging processing mechanism, failing to establish a closed-loop mechanism for signal perception and dynamic decision-making. This affects the real-time nature and accuracy of pricing decisions, making the system unable to adapt to the rapidly changing market environment.

[0027] For example, when a tourist city suddenly announces the hosting of a large-scale sporting event, traditional systems can only make predictions based on historical data from the same period, failing to obtain crucial information such as the event's scale, expected number of participants, and geographical distribution in a timely manner. Meanwhile, sudden regional weather changes (such as sudden high temperatures or torrential rains) and abnormal traffic flow around hotels further exacerbate demand uncertainty, but the system lacks the ability to dynamically correlate meteorological and traffic data. In this scenario, hotel pricing fails to adjust to actual demand fluctuations, leading to lower-than-expected revenue during periods of high demand or increased vacancy rates during periods of low demand.

[0028] Specifically, the system fails to identify the contribution of event types to demand, and also fails to quantify the difference between discomfort in the customer's origin and the hotel's location, resulting in a mismatch between pricing strategies and actual market demand. If these issues are not addressed, the system will continue to be unable to adapt to dynamic market changes, leading to a decline in the real-time nature and accuracy of pricing decisions. Consequently, the hotel will be at a disadvantage in market competition, negatively impacting long-term revenue stability and inventory management efficiency, and hindering the achievement of revenue optimization goals.

[0029] Therefore, such as Figure 1 As shown, this application provides a big data-driven intelligent system for dynamic hotel pricing and revenue optimization, including the following modules:

[0030] The data acquisition module is used to synchronously acquire urban event data, regional meteorological data, and hotel traffic data within a preset monitoring period;

[0031] The data evaluation module is used to extract event scores and weather scores based on the city event data and regional meteorological data, respectively, and to extract a traffic score based on the weather scores combined with hotel traffic data.

[0032] The data fusion module is used to sort the various scores to obtain the first demand index, the second demand index, and the third demand index, and to obtain the dynamic demand index by combining the preset fusion weights of the corresponding order.

[0033] The pricing generation module is used to input the dynamic demand index and the first demand index into a preset pricing mapping model to output a pricing adjustment coefficient, and generate a corresponding preliminary price by combining the basic price of the target room type.

[0034] The pricing correction module is used to obtain the competitor pricing of the target room type within a preset monitoring period, and to correct the preliminary pricing by combining it with a preset sensitivity coefficient corresponding to the score of the first demand index to obtain its final pricing.

[0035] The feedback optimization module is used to obtain the order volatility index of the target room type within a preset time window after the final pricing is executed, and to update the preset fusion weight or preset sensitivity coefficient according to the index until the order volatility index that meets the preset output conditions is obtained.

[0036] This application introduces a data acquisition module to simultaneously acquire multi-source implicit demand signals, including urban event data, regional meteorological data, and hotel traffic data, thereby achieving market demand perception. This contrasts with existing technologies that rely on historical occupancy rates and competitor pricing data. For example, by analyzing urban event data related to conferences, the system can predict peak demand in advance, rather than adjusting prices only after a surge in orders.

[0037] Furthermore, the data evaluation module transforms this data into event scores, weather scores, and traffic scores, enabling the measurement and comparison of different data types. The data fusion module then generates a dynamic demand index by sorting these scores and combining them with preset fusion weights. This multi-dimensional scoring fusion mechanism allows the system to characterize market demand and avoids the potential biases that indicators may introduce. For example, even during periods of high temperatures and thunderstorms, if there are meetings scheduled, the system can consider weather and traffic scores to avoid setting prices based on optimism, which could lead to customer churn.

[0038] The pricing generation and pricing correction modules work together to not only generate an initial price based on a dynamic demand index, but also adjust the price using competitor pricing and a preset sensitivity coefficient. This ensures that the pricing reflects both the system's assessment of market demand and the competitive landscape, guaranteeing price competitiveness. For example, with an initial price of 600 yuan, the system will refer to competitor pricing and adjust it based on market sensitivity to avoid prices being too high or too low, thus maintaining market share while ensuring revenue.

[0039] The feedback optimization module constructs a closed-loop learning mechanism. By monitoring the order volatility index, the system can evaluate the effectiveness of the pricing strategy and automatically adjust the preset fusion weights or preset sensitivity coefficients based on the feedback results. This adaptive learning capability allows the system to evolve with changes in the market environment and optimize its pricing strategy. For example, if the initially set fusion weights cause order volatility, the system can automatically adjust the weights, making the pricing strategy robust. This overcomes the limitations of traditional systems' static pricing or manual experience adjustments, improving pricing accuracy and revenue stability.

[0040] In summary, the system in this embodiment achieves a shift from historical data-driven to perception-driven approaches through a closed loop of "signal perception - intelligent fusion - game theory decision-making - feedback evolution." This enables it to capture and characterize demand opportunities, resulting in accurate and agile pricing decisions. By integrating multi-dimensional signals and introducing market competition, it balances demand capture with market competitiveness. The closed-loop feedback mechanism ensures the system evolves with market changes, improving hotel profitability and market responsiveness, and providing hotel managers with revenue optimization tools.

[0041] It should be further explained that, in the specific implementation process, the process of simultaneously acquiring urban event data, regional meteorological data, and hotel traffic data within the preset monitoring period includes:

[0042] The urban event data refers to an ordered collection of information on various public activities and events that may cause fluctuations in local accommodation demand, collected within a preset monitoring period from the city where the hotel is located. Its core function is to capture hidden peak season signals outside of traditional holidays and seasonal patterns, providing key input for predicting short-term, localized high demand.

[0043] The city event data includes: event title and description text, event type (such as professional exhibitions, academic conferences, large-scale competitions, cultural performances, special events at famous scenic spots, important examinations, large-scale corporate recruitment fairs, etc.), officially released estimated scale (such as the number of exhibitors, the number of visitors, and the number of examinees), the precise geographical coordinates (latitude and longitude) of the event location, and the specific start and end dates and key time points of the event.

[0044] Data sources typically include: local government information disclosure platforms (such as exhibition / event registration announcements on the official websites of commerce bureaus and culture and tourism bureaus), official websites and schedule pages of major convention and exhibition centers, ticketing and performance plans of sports venues and theaters, annual examination schedules published by education examination institutions, and campus recruitment or social recruitment announcement pages of large enterprises. Data acquisition is achieved by configuring targeted web crawler programs or calling the open API interfaces of relevant websites.

[0045] The regional meteorological data refers to a set of meteorological element observations or forecasts covering the hotel's location and core source cities (usually the top 5-10 cities with the highest booking volume) identified based on historical order data within a preset monitoring period. Its core function is to capture immediate travel demand driven by climate adaptation motivations such as seeking refuge from heat, cold, and smog by quantifying the climate differences between the source and destination.

[0046] The regional meteorological data includes: temperature, relative humidity, precipitation, wind speed and direction, air quality index (AQI), ultraviolet radiation index, perceived temperature, and weather forecast data for the next 24-72 hours. Data sources are typically standardized API interfaces provided by commercial meteorological service providers (such as service companies under the China Meteorological Administration), or reanalysis datasets obtained from authoritative institutions such as the National Meteorological Science Data Center.

[0047] The hotel traffic data refers to a dynamic data set reflecting the real-time traffic conditions and efficiency between the hotel and key transportation hubs and core functional areas inside and outside the city within a preset monitoring period. Its core function is to perceive the instantaneous changes (improvement or deterioration) in the city's micro-traffic environment and assess its immediate impact on the hotel's accessibility and accommodation attractiveness.

[0048] The hotel's traffic data includes: real-time traffic congestion index (or traffic volume) of major roads surrounding the hotel, real-time arrival / delay information of subway and bus lines, and real-time estimated travel time and distance by driving / public transportation from the hotel to preset key nodes (usually including at least: the nearest airport terminal, high-speed rail station, the city center's core business district, and a representative popular scenic spot). The data source is mainly real-time traffic and route planning APIs provided by major map service providers such as Gaode Maps and Baidu Maps.

[0049] It should be further explained that, in the specific implementation process, the extraction of event scores and meteorological scores includes: performing natural language processing on the urban event data to obtain its event type and estimated number of participants. and the coordinates of the event venue, the event score ;in, This is a preset type coefficient for the event type. The distance between the coordinates of the event venue and the hotel is expressed in Euclidean form. The preset distance threshold;

[0050] The discomfort level was obtained based on regional meteorological data corresponding to the hotel's location and several source locations of customers. This discomfort level was obtained by weighted summation of different meteorological indicators, representing the inhospitability of the corresponding area. The meteorological score... ,in, This represents the average level of discomfort in each source city. The level of discomfort in the hotel's location.

[0051] Specifically, natural language processing (NLP) of urban event data aims to automatically identify and extract key elements from unstructured text information, such as the nature of the event (event type), its scale (expected number of participants S), and its location (location coordinates). This enables the system to transform raw, difficult-to-analyze text data into structured, quantifiable information, laying the foundation for subsequent event scoring calculations. This process can employ rule-based methods, using preset keywords and patterns to identify event types, extract numerical information as the expected number of participants S, and utilize geocoding technology to convert location descriptions into precise latitude and longitude coordinates. Alternatively, advanced machine learning models, such as deep learning networks, can be trained on large amounts of labeled data to extract the necessary information more intelligently and accurately from complex urban event data.

[0052] The event score is a numerical value that quantifies the impact of urban events on hotel demand. Its formula comprehensively considers the type, scale, and geographical distance of the event from the hotel. The preset type coefficient is used to reflect the inherent differences in the impact of different types of events (such as international conferences, large-scale concerts, local festivals, etc.) on hotel demand. The term represents the impact of the expected number of participants S on demand. Generally, the more participants there are, the greater the increase in demand, but its marginal effect will gradually decrease. The term reflects the distance decay effect, meaning that the greater the Euclidean distance between the event location and the hotel coordinates, the smaller its impact on hotel demand. A preset distance threshold is used to adjust the rate of distance decay. This formula can be implemented based on historical data analysis, determined and adjusted through regression models or expert experience. and The numerical value is used to ensure the accuracy of the score. The above is the method for obtaining the score of a single event. If there are multiple events, the average value is taken as the event score for the corresponding period.

[0053] The discomfort level is obtained by analyzing regional meteorological data corresponding to the hotel's location and several source locations. This aims to indirectly reflect people's travel intentions by comprehensively assessing the impact of meteorological conditions on human comfort. Discomfort level is a comprehensive index obtained by weighted summation of multiple meteorological indicators (such as temperature, humidity, wind speed, and light intensity), and its value indicates the degree of inhospitability of the corresponding area. Mature models such as the Temperature and Humidity Index (THI) or Effective Temperature (ET) can be used to calculate discomfort level, and the weights of each meteorological indicator can be adjusted according to actual conditions. Furthermore, more complex models can be constructed to incorporate factors such as the Air Quality Index (AQI) and precipitation probability to provide a more comprehensive assessment of discomfort level.

[0054] The meteorological score This represents the average level of discomfort in each source city. The system measures the discomfort level of the hotel's location. Its formula quantifies the impact of weather conditions on hotel demand by comparing the discomfort levels of the source city and the hotel's location. When the discomfort level in the source city is significantly higher than that in the hotel's location, it indicates that residents of the source city have a stronger incentive to travel to the more pleasant climate of the hotel location, potentially increasing hotel demand; conversely, if the hotel's location has unfavorable weather conditions, it may suppress demand. The system can periodically acquire weather data for the hotel's location and major source cities, calculate the discomfort levels for each, and then obtain a weather score, thus providing a quantitative basis for dynamic hotel pricing based on weather factors.

[0055] Through the aforementioned technical solutions, this application can accurately extract key information from unstructured urban event data using natural language processing technology, and precisely calculate event scores using quantitative formulas by combining event type, scale, and distance. By comparing the differences in weather discomfort between the hotel's location and the customer's origin, the impact of weather conditions on potential customers' travel intentions can be objectively reflected, thereby obtaining more insightful weather scores. These refined, data-driven scoring extraction methods significantly improve the accuracy and comprehensiveness of the data evaluation module's quantification of the impact of external environmental factors on hotel demand, providing more reliable and accurate input for subsequent dynamic demand index calculations, and effectively improving the accuracy of hotel dynamic pricing and revenue optimization.

[0056] It should be further explained that, in the specific implementation process, the process of extracting traffic scores includes: obtaining the traffic flow in a preset area centered on the hotel coordinates within the current preset monitoring period based on the hotel traffic data. And obtain the average traffic flow of the preset area over the most recent n preset monitoring periods. Combined with the meteorological score within the current preset monitoring period Get traffic rating ,in, This is a preset adjustment factor.

[0057] The preset area centered on the hotel's coordinates refers to a specific range defined around the hotel's geographical location, used to focus on analyzing local traffic conditions directly related to the hotel. This area can be a circular area with the hotel as the center and a preset radius, such as a range with a radius of 1 kilometer or 2 kilometers; or it can be an arbitrary polygonal area defined based on factors such as administrative divisions, business district boundaries, or major traffic arteries.

[0058] Traffic flow within the current preset monitoring period refers to the volume or density of vehicles passing through the preset area during the currently ongoing data monitoring time period. For example, if the preset monitoring period is one hour, it can represent the total number of vehicles passing through the area in the current hour, or the average vehicle density within the area. The average traffic flow over the most recent n preset monitoring periods refers to the average traffic flow in the preset area over the n consecutive identical monitoring periods prior to the current monitoring period. This average serves as a benchmark to measure the relative level of the current traffic flow. The preset adjustment factor is a parameter used to adjust the degree of influence of weather scores on traffic scores. This factor can be obtained based on historical data analysis, expert experience, or training through machine learning models to reflect the sensitivity of traffic flow to different weather conditions.

[0059] For example, during inclement weather, even if traffic volume is relatively high, its positive impact on hotel demand may be appropriately reduced considering the weather factor. Conversely, during pleasant weather, relatively high traffic volume may be perceived as having a stronger positive impact on hotel demand. This combined approach allows traffic scoring to more accurately capture regional activity influenced by both traffic and weather, providing more refined input for subsequent dynamic demand index calculations and ultimately improving pricing accuracy.

[0060] Through the aforementioned technical solutions, the system can more precisely capture the real-time activity level of the area surrounding the hotel. By comparing current traffic flow with historical averages and further refining it with weather scores, the resulting traffic score more accurately reflects the potential demand influenced by both traffic and weather. This allows the system to fully consider local, real-time environmental factors when assessing the hotel's dynamic demand index, thus avoiding assessment biases caused by relying solely on macro-level events and regional weather data. Ultimately, this more accurate demand assessment helps generate preliminary pricing that better reflects market realities, effectively improving the accuracy of dynamic hotel pricing and revenue optimization.

[0061] It should be further explained that, in the specific implementation process, the process of obtaining the dynamic demand index includes: scoring events within the same preset monitoring period. Weather score Traffic scoring After normalization, the values ​​are sorted in descending order. The value ranked first is taken as the first demand index, the value ranked second as the second demand index, and the value ranked third as the third demand index; this constitutes the first demand index. Second demand index Third demand index Set the first fusion weight respectively Second fusion weight Third fusion weight , The dynamic demand index The preset fusion weights include a first fusion weight, a second fusion weight, and a third fusion weight.

[0062] Normalization refers to converting event scores, weather scores, and traffic scores with different dimensions or numerical ranges into a unified numerical interval [0, 1] to eliminate the impact of dimensional differences on subsequent calculations. This can be achieved through min-max normalization, Z-score standardization, or other suitable mathematical methods. Descending order ranking means that after normalization, these three scores are arranged from highest to lowest according to their numerical values ​​to identify the factors that have the greatest impact on hotel demand within the current preset monitoring period. The first demand index, second demand index, and third demand index are determined based on the ranking results, representing the strongest, second strongest, and third strongest factors influencing hotel demand within the current preset monitoring period, respectively. This method can intuitively reflect which external factors play a dominant role in hotel demand within a specific time period.

[0063] The fusion weights are used to quantify the relative importance of each demand index in the calculation of the dynamic demand index. By setting different weights, it can be ensured that factors with a greater impact on hotel demand have a higher proportion in the final dynamic demand index. The design explicitly states that demand indices ranked higher (i.e., factors with greater influence) will receive higher weights, thus contributing more significantly to the calculation of the dynamic demand index and making it more accurate in reflecting the impact of dominant factors. These weights can be pre-defined empirical values ​​or obtained through training a machine learning model. Dynamic Demand Index It is the result of a weighted sum of various demand indices and their corresponding weights. This weighted summation method can comprehensively consider the influence of different factors and adjust them according to their relative importance, thereby generating a comprehensive indicator that can fully and accurately reflect the demand of the hotel market.

[0064] Through the above technical solution, this application can uniformly normalize event scores, weather scores, and traffic scores from different dimensions and with different dimensions, and rank them in descending order according to their impact on hotel demand, thereby accurately identifying the dominant factors of current market demand. Based on this, by assigning higher fusion weights to the top-ranked demand indices and performing a weighted summation, a dynamic demand index is obtained. This approach allows for a more accurate and objective reflection of the hotel's true market demand. It avoids the information distortion that can result from simple averaging or equal-weighted summation, enabling the system to more sensitively capture changing market demand trends. This provides a more refined and reliable basis for subsequent pricing decisions, effectively improving the accuracy of the hotel's dynamic pricing and its revenue optimization capabilities.

[0065] It should be further explained that, in the specific implementation process, the pricing mapping model is a preset function that takes the dynamic demand index and the first demand index as inputs and the pricing adjustment coefficient as output; the preset demand range corresponding to the first demand index is obtained. The dynamic demand index Compared with the preset demand range, if Then output the pricing adjustment coefficient. ,like Then output the pricing adjustment coefficient. ,like Then output the pricing adjustment coefficient. , ; Obtain the base price of the target room type within the current preset monitoring period, and multiply it by the price adjustment coefficient output by the current preset monitoring period as the preliminary price.

[0066] The pricing mapping model can be a mathematical function, a lookup table, or a rule-based decision system. Its core function is to output a suitable pricing adjustment coefficient based on the input dynamic demand index and the first demand index. This model can be implemented in various forms. For example, it can be a polynomial function fitted from historical data; it can be a piecewise function that outputs different adjustment coefficients based on different input intervals; or it can be a machine learning model, such as a support vector machine or neural network, which learns the complex relationship between input and output through training.

[0067] The preset demand range The purpose of obtaining this range is to provide a reference range for the dynamic demand index, in order to determine the relative level of current demand. This range can be obtained through statistical analysis of historical data of the first demand index; for example, the 25th percentile of the first demand index in historical data can be set as... The 75th percentile is set as Alternatively, settings can be made based on expert experience or market research results to reflect the hotel's understanding of changes in demand under specific market conditions.

[0068] The pricing adjustment coefficient is a key parameter used to adjust the base price, and its magnitude is related to... This reflects a strategy of adjusting prices differently based on varying levels of demand. When the dynamic demand index... Higher than When this occurs, it indicates strong market demand, and a larger pricing adjustment coefficient should be applied. To increase prices and maximize profits; when the dynamic demand index In and When the time frame is between these two points, it indicates that market demand is at a normal level, and at this time, an appropriate pricing adjustment coefficient should be output. To maintain stable prices and occupancy rates; when the dynamic demand index Below When this indicates weak market demand, a smaller pricing adjustment coefficient should be applied. In order to lower prices, stimulate demand, and avoid excessively high vacancy rates, the base pricing refers to the benchmark price of the target housing type under the influence of no external demand fluctuations.

[0069] Through the aforementioned technical solution, this application can generate preliminary pricing more precisely based on differences in market demand levels. By comparing the dynamic demand index with a preset demand range and outputting different pricing adjustment coefficients based on the comparison results, the preliminary pricing can more accurately reflect the current market supply and demand relationship. This helps hotels seize opportunities to increase revenue when demand is high and stimulate consumption through price adjustments when demand is weak, thereby avoiding revenue losses or market share declines caused by overly simplistic pricing strategies, and improving pricing accuracy and market responsiveness.

[0070] It should be further clarified that, in the specific implementation process, the "competitive pricing" refers to the real-time pricing of the same room type as the target room type obtained from several hotels determined through preset screening rules; the obtained competitive pricing is sorted according to its numerical value, and the median of these prices is taken as the median competitive pricing. A preset sensitivity coefficient is obtained based on the first demand index to determine its corresponding score. Obtain the upper limit of price fluctuation and price fluctuation floor ;

[0071] The initial pricing Compared with the upper and lower limits of price fluctuation, if The initial pricing will then be set. As the final pricing, if Then the upper limit of price fluctuation will be set. As the final pricing, if Then the lower limit of price fluctuation will be set. As the final price.

[0072] Competitive pricing refers to the real-time pricing of the same room type as the target room type obtained from several hotels selected through preset filtering rules. It reflects the pricing information offered by competitors in the market with similar positioning, service levels, and target customer groups. Competitive pricing can be achieved through web scraping technology, periodically or in real-time capturing hotel room price data from publicly available channels such as OTA (Online Travel Agency) platforms and competitor websites, and then filtering and matching this data based on factors such as hotel location, star rating, and brand. Alternatively, it can be obtained by partnering with third-party data service providers, whose market competitive pricing data is typically cleaned and standardized to more accurately reflect the market competition landscape.

[0073] Preset filtering rules are a set of standards used to determine which hotels are considered competitors of the target hotel. These rules ensure that the selected competitor hotels are comparable to the target hotel in terms of location, star rating, brand positioning, room type similarity, and service facilities. Rules can include geographical distance restrictions, star rating matching, brand type, and room type characteristics. Rules can also be intelligently analyzed using unstructured data such as user reviews, market share, and historical occupancy rates to dynamically adjust the competitor hotel list and more accurately reflect the market competition landscape. The median competitor pricing is calculated by sorting the obtained competitor prices by numerical value and taking the middle value. After obtaining all competitor pricing data, these data are stored in a list, which is then sorted in ascending or descending order. If the number of list elements is odd, the median is the middle element; if it is even, the median is the average of the two middle elements.

[0074] The preset sensitivity coefficient, a parameter derived from the first demand index, quantifies the hotel's responsiveness to changes in market demand, or in other words, the extent to which the hotel is willing to adjust its prices in the face of competitive market pricing. This coefficient reflects the hotel's strategic inclination towards price competition under different levels of market demand. A mapping relationship between the first demand index and the hotel's price adjustment range can be established through historical data analysis. It can also be set and dynamically adjusted using expert experience or machine learning models. The upper and lower limits of price fluctuation are price range boundaries calculated based on the median competitor pricing and the preset sensitivity coefficient. This range defines a reasonable range for the target hotel's final pricing, ensuring that the price is neither significantly higher than the mainstream market price, leading to a decrease in competitiveness, nor significantly lower than the mainstream market price, resulting in revenue loss.

[0075] The preliminary price is an initial price generated by the system after considering various demand factors such as city events, weather, and traffic, combined with the base price of the target room type. It reflects the theoretically optimal price based on market demand forecasts. The final price is the final selling price after adjustments for competitor pricing. It is the actual publicly announced price determined by the system for the target room type after comprehensively considering market demand, the hotel's own situation, and competitor pricing. The final price is determined by comparing the preliminary price with calculated upper and lower price fluctuation limits to ensure that the final price falls within a reasonable competitive range. This adjustment mechanism makes the hotel's pricing strategy more flexible, allowing adjustments based on real-time changes in market competition, thereby maximizing profits while maintaining competitiveness.

[0076] Through the aforementioned technical solution, this application effectively addresses the problem of hotel pricing relying solely on internal data and demand forecasting while neglecting external market competition. By introducing real-time monitoring and median calculation of competitor pricing, and combining this with a preset sensitivity coefficient to dynamically set the price fluctuation range, the system ensures that the final price always remains within a reasonable range that responds to market demand while also considering the competitive landscape. This avoids the risk of losing customers due to excessively high pricing or suffering revenue losses due to excessively low pricing, thereby significantly enhancing the hotel's market competitiveness, optimizing revenue management, and enabling the hotel to maintain a flexible and advantageous pricing strategy in a fiercely competitive market.

[0077] It should be further explained that, in the specific implementation process, the process of updating the preset fusion weight includes: obtaining the number of orders for the target room type within the current preset time window after the execution of the final pricing and within multiple consecutive preset time windows before the execution of the final pricing, and obtaining the variance of the number of orders within the current preset time window as the order fluctuation index;

[0078] When the order volatility index is greater than a preset first volatility threshold At that time, an iterative update of the preset fusion weights is triggered, obtaining the first fusion weight after a single update. and the second fusion weight and the third fusion weight ;

[0079] The order volatility index is obtained after each update until it meets the preset output conditions. Then, the first, second, and third fusion weights at this point are applied to the next preset monitoring period. These are the first fusion weight, second fusion weight, and third fusion weight before the corresponding single update. The minimum allowed first fusion weight, Adjust the step size of the preset weights.

[0080] The order volatility index is a key indicator for measuring the stability of order volume for a target room type. It is calculated by obtaining the order volume for the target room type within the current preset time window after the final pricing is implemented, as well as within several consecutive preset time windows before the final pricing is implemented, and then calculating the variance of the order volume within the current preset time window. This index reflects the market's responsiveness to the current pricing strategy; a higher order volatility index indicates a potential significant deviation between the pricing strategy and actual market demand, leading to unstable order volume.

[0081] The preset first fluctuation threshold is a pre-defined value used to determine whether the order fluctuation index is within an acceptable range. When the order fluctuation index exceeds this threshold, it indicates that order fluctuations are too large, requiring adjustments to the pricing strategy. Iterative updates refer to the process by which the system repeatedly adjusts the preset fusion weights based on feedback information. When the order fluctuation index exceeds the preset first fluctuation threshold, the system triggers this update mechanism. Each update adjusts the fusion weights according to a specific algorithm and then re-evaluates its effectiveness until the preset optimization goal is achieved. In addition to the specific formula given in this application, iterative updates can also be implemented using gradient descent-based optimization algorithms, reinforcement learning algorithms, or heuristic rules.

[0082] First fusion weight Second fusion weight Third fusion weight These are the new fusion weights obtained after a single update. These weights are used to redistribute the influence of the first, second, and third demand indices in subsequent dynamic demand index calculations. Ensuring the first fusion weight... The weight will not fall below the preset minimum first fusion weight, and will decrease incrementally with a preset weight reduction step size. Second fusion weight and the third fusion weight Then, based on its relative proportion before the update, it is increased proportionally to compensate for the first fusion weight. The reduction in weights ensures that the sum of all fusion weights remains constant or nearly constant, thereby maintaining the rationality of the dynamic demand index.

[0083] The minimum allowed first fusion weight sets a lower limit for the first fusion weight. Setting this minimum prevents the first demand index from being excessively weakened during iterations, ensuring it maintains a certain weight and influence in the dynamic demand index calculation. The preset weight reduction step size determines the extent to which the first fusion weight is reduced in each iteration. The size of this step size affects the speed and stability of the iteration update; a smaller step size can make the update process smoother but may require more iterations; a larger step size may speed up convergence but may also lead to instability in the update process.

[0084] In this way, the system can dynamically adjust the influence of different demand indices in the calculation of the dynamic demand index. For example, if order fluctuations are too large, it may mean that the current pricing strategy is overly reliant on a certain demand factor, or that its weighting is unreasonable. By reducing the first fusion weight and correspondingly increasing other fusion weights, the system can make the calculation of the dynamic demand index more balanced and robust, thereby generating pricing that better reflects the actual market situation. This iterative update process continues, and the order fluctuation index is re-evaluated after each update until it meets the preset output conditions, i.e., the order fluctuation index reaches an acceptable stable range. Once the conditions are met, the fusion weights are locked and applied to the next preset monitoring cycle, thus forming a closed-loop feedback optimization mechanism. This allows the system to continuously learn and adapt to market changes, ensuring the effectiveness of the pricing strategy and the optimization of returns.

[0085] Through the above technical solution, this application effectively addresses the problem of excessive hotel order fluctuations caused by inaccurate preset fusion weights. By introducing an order volatility index as a feedback mechanism and dynamically adjusting the preset fusion weights based on its magnitude, the system can achieve adaptive learning and optimization in response to market demand. When the order volatility index exceeds a preset first volatility threshold, the system can promptly trigger iterative updates to the fusion weights, making the calculation of the dynamic demand index closer to actual market conditions, thereby generating more stable and reasonable preliminary and final pricing. This feedback optimization mechanism enables the hotel's pricing strategy to continuously adapt to market changes, effectively reducing order volatility risk, improving the stability and predictability of revenue, and ultimately achieving continuous optimization of hotel revenue.

[0086] It should be further explained that, in the specific implementation process, the process of updating the preset sensitivity coefficient includes: when the order volatility index is less than the preset second volatility threshold... At that time, the preset sensitivity coefficient corresponding to the score of the first demand index is triggered for iterative update, and the preset sensitivity coefficient after a single update is obtained. ,in, , To correspond to the preset sensitivity coefficient before a single update, The maximum allowed preset sensitivity coefficient. Adjust the step size for the preset coefficient;

[0087] The order volatility index is obtained after each update until it meets a preset output condition. Then, the preset sensitivity coefficient of the score corresponding to the first demand index at that time is applied to the next preset monitoring period. The preset output condition is that the corresponding order volatility index is greater than or equal to a preset second volatility threshold. And less than or equal to the preset first fluctuation threshold .

[0088] Among them, when the order volatility index is less than the preset second volatility threshold This indicates that after implementing the final pricing, the number of orders for the target room type changed little, and the market response was lukewarm. This may mean that the current pricing strategy was too conservative and failed to fully stimulate market demand or effectively utilize market potential. This condition can be determined by continuously monitoring the order volatility index through the system and comparing it with a preset second volatility threshold. The system performs real-time comparisons to determine the sensitivity. Once the order volatility index falls below this threshold, the subsequent sensitivity coefficient update mechanism is triggered.

[0089] Triggering an iterative update of the preset sensitivity coefficient corresponding to the score of the first demand index means that the system automatically initiates an adjustment process for the preset sensitivity coefficient when it detects that the order volatility index is too low. The preset sensitivity coefficient is used in the pricing correction module to determine the upper and lower limits of price fluctuations in conjunction with the median pricing of competing products, thereby correcting the initial pricing. Its iterative update aims to gradually adjust the responsiveness of the pricing strategy to market changes, making it more flexible and improving order volatility. This triggering mechanism can be executed automatically by a background algorithm, such as through a scheduled task or event-driven approach.

[0090] in, This is the currently used preset sensitivity coefficient. This is a preset step size for adjusting the sensitivity coefficient, representing the amount by which the sensitivity coefficient increases with each update. It's necessary to ensure that the updated sensitivity coefficient does not exceed the preset maximum allowable value, thereby preventing excessively high sensitivity coefficients that could lead to overly aggressive or unstable pricing. For example, It can be a fixed small value, such as 0.01 or 0.005, to obtain the order volatility index after each update until it meets the preset output conditions. This means that after each update of the preset sensitivity coefficient and application of a new pricing strategy, the system will monitor the order volatility index of the target room type again. This process will continue to form a feedback loop until the order volatility index reaches the preset ideal range.

[0091] The preset output condition is that the corresponding order volatility index is greater than or equal to a preset second volatility threshold and less than or equal to a preset first volatility threshold. This condition defines the termination criterion for the iterative update process. It means that when the order volatility index is in a range that is neither too low (avoiding conservative pricing) nor too high (avoiding aggressive pricing), the sensitivity coefficient update stops. This range ensures that the pricing strategy can fully utilize market opportunities while maintaining relative stability and avoiding excessive volatility. For example, It can be set to 0.05. It can be set to 0.15, indicating that an order volatility index between 5% and 15% is considered ideal.

[0092] Through the aforementioned technical solution, when the order volatility index is too low, the system can automatically identify and trigger iterative updates of the preset sensitivity coefficient. This mechanism allows the pricing strategy to gradually adjust from an overly conservative state to a more proactive and market-responsive one, effectively avoiding missing potential revenue growth opportunities due to conservative pricing. Combined with the mechanism for handling excessively high order volatility, it constructs a more comprehensive and robust dynamic pricing optimization system, ensuring that hotel pricing is always within a dynamic equilibrium range that maximizes revenue.

[0093] In another implementation, such as Figure 2 As shown, this application also provides a big data-driven intelligent method for dynamic hotel pricing and revenue optimization, including the following steps:

[0094] Simultaneously acquire urban event data, regional meteorological data, and hotel traffic data within a preset monitoring period;

[0095] Based on the city event data and regional meteorological data, event scores and meteorological scores are extracted respectively. Based on the meteorological scores and hotel traffic data, a traffic score is extracted.

[0096] The scores are ranked to obtain the first demand index, the second demand index, and the third demand index. The dynamic demand index is obtained by combining the preset fusion weights of the corresponding order.

[0097] The dynamic demand index and the first demand index are input into a preset pricing mapping model to output a pricing adjustment coefficient, which is then combined with the base price of the target room type to generate the corresponding preliminary price.

[0098] Obtain the competitor pricing of the target room type within a preset monitoring period, and adjust the preliminary pricing by combining it with the preset sensitivity coefficient of the score corresponding to the first demand index to obtain its final pricing;

[0099] Obtain the order volatility index of the target room type within a preset time window after the final pricing is executed, and update the preset fusion weight or preset sensitivity coefficient according to it until the order volatility index that meets the preset output conditions is obtained.

[0100] Through the aforementioned technical solutions, the system can capture and quantify unconventional demand opportunities, enabling more accurate and agile pricing decisions. For example, before a large conference, the system identifies incremental demand through event scoring, and dynamically adjusts the fusion weights to generate a dynamic demand index by combining the decrease in meteorological scores due to high temperatures and the fluctuations in traffic scores caused by traffic congestion, thus avoiding pricing deviations caused by a single signal. Simultaneously, through the synergistic effect of the median competitor pricing and a preset sensitivity coefficient, the system can appropriately increase prices during peak demand periods to capture revenue while preventing excessively high prices from causing customer churn, effectively balancing revenue maximization and market competitiveness. The closed-loop feedback mechanism uses the order fluctuation index as an optimization basis, automatically adjusting the preset fusion weights when order volume fluctuates abnormally, such as reducing the weight of dominant signals, allowing the system to continuously evolve with changes in the market environment. This multi-dimensional signal fusion and dynamic game-theoretic decision-making mechanism significantly improves the hotel's responsiveness to market demand and pricing accuracy, optimizes long-term revenue levels, and provides hotel managers with an intelligent revenue optimization tool.

[0101] In another embodiment, this application also provides a computer storage medium storing computer-executable instructions, which, when executed, implement the big data-driven intelligent system for dynamic hotel pricing and revenue optimization.

[0102] The above embodiments are only used to illustrate the technical methods of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of this application without departing from the spirit and scope of the technical methods of this application.

Claims

1. A big data-driven intelligent system for dynamic hotel pricing and revenue optimization, characterized in that: Includes the following modules: The data acquisition module is used to synchronously acquire urban event data, regional meteorological data, and hotel traffic data within a preset monitoring period; The data evaluation module is used to extract event scores and weather scores based on the city event data and regional meteorological data, respectively, and to extract a traffic score based on the weather scores combined with hotel traffic data. The data fusion module is used to sort the various scores to obtain the first demand index, the second demand index, and the third demand index, and to obtain the dynamic demand index by combining the preset fusion weights of the corresponding order. The pricing generation module is used to input the dynamic demand index and the first demand index into a preset pricing mapping model to output a pricing adjustment coefficient, and generate a corresponding preliminary price by combining the basic price of the target room type. The pricing correction module is used to obtain the competitor pricing of the target room type within a preset monitoring period, and to correct the preliminary pricing by combining it with a preset sensitivity coefficient corresponding to the score of the first demand index to obtain its final pricing. The feedback optimization module is used to obtain the order volatility index of the target room type within a preset time window after the final pricing is executed, and to update the preset fusion weight or preset sensitivity coefficient according to the index until the order volatility index that meets the preset output conditions is obtained.

2. The big data-driven intelligent system for dynamic hotel pricing and revenue optimization according to claim 1, characterized in that, The process of extracting event scores and weather scores includes: Natural language processing is performed on the city event data to obtain its event type and estimated number of participants. and the coordinates of the event venue, the event score ; in, This is a preset type coefficient for the event type. The distance between the coordinates of the event venue and the hotel is expressed in Euclidean form. The preset distance threshold; The discomfort level was obtained based on regional meteorological data corresponding to the hotel's location and several customer source locations. The discomfort level was obtained by weighted summation of different meteorological indicators, which represents the degree of inhospitability of the corresponding area. The meteorological score ,in, This represents the average level of discomfort in each source city. The level of discomfort in the hotel's location.

3. The big data-driven intelligent system for dynamic hotel pricing and revenue optimization according to claim 2, characterized in that, The process of extracting traffic scores includes: Based on the hotel traffic data, the traffic flow in a preset area centered on the hotel coordinates within the current preset monitoring period is obtained. And obtain the average traffic flow of the preset area over the most recent n preset monitoring periods. ; Combined with meteorological scores within the current preset monitoring period Get traffic rating ,in, This is a preset adjustment factor.

4. The big data-driven intelligent system for dynamic hotel pricing and revenue optimization according to claim 3, characterized in that, The process of obtaining the dynamic demand index includes: Scoring events within the same preset monitoring period Weather score Traffic scoring After normalization, the values ​​are sorted in descending order. The value ranked first is taken as the first demand index, the value ranked second is taken as the second demand index, and the value ranked third is taken as the third demand index. The first demand index Second demand index Third demand index Set the first fusion weight respectively Second fusion weight Third fusion weight , The dynamic demand index The preset fusion weights include a first fusion weight, a second fusion weight, and a third fusion weight.

5. The big data-driven intelligent system for dynamic hotel pricing and revenue optimization according to claim 1, characterized in that, The pricing mapping model is a preset function that takes the dynamic demand index and the first demand index as inputs and the pricing adjustment coefficient as output. The preset demand range is obtained based on the first demand index and its corresponding score. The dynamic demand index Compared with the preset demand range, if Then output the pricing adjustment coefficient. ,like Then output the pricing adjustment coefficient. ,like Then output the pricing adjustment coefficient. , ; Obtain the base price of the target room type within the current preset monitoring period, and multiply it by the price adjustment coefficient output in the current preset monitoring period as the preliminary price.

6. The big data-driven intelligent system for dynamic hotel pricing and revenue optimization according to claim 1, characterized in that, The competitor pricing refers to the real-time pricing of the same room type as the target room type obtained from several hotels determined through preset screening rules; Sort the obtained competitor pricing data by numerical value, and use the median of these medians as the median competitor pricing. A preset sensitivity coefficient is obtained based on the first demand index to determine its corresponding score. Obtain the upper limit of price fluctuation and price fluctuation floor ; The initial pricing Compared with the upper and lower limits of price fluctuation, if The initial pricing will then be set. As the final pricing, if Then the upper limit of price fluctuation will be set. As the final pricing, if Then the lower limit of price fluctuation will be set. As the final price.

7. The big data-driven intelligent system for dynamic hotel pricing and revenue optimization according to claim 4, characterized in that, The process of updating the preset fusion weights includes: Obtain the number of orders for the target room type within the current preset time window after the final pricing is implemented and within multiple consecutive preset time windows before the final pricing is implemented; obtain the variance of the number of orders within the current preset time window as the order volatility index. When the order volatility index is greater than a preset first volatility threshold At that time, an iterative update of the preset fusion weights is triggered, obtaining the first fusion weight after a single update. and the second fusion weight and the third fusion weight ; The order volatility index is obtained after each update until it meets the preset output conditions. Then, the first, second, and third fusion weights at this point are applied to the next preset monitoring period. These are the first fusion weight, second fusion weight, and third fusion weight before the corresponding single update. The minimum allowed first fusion weight, Adjust the step size for the preset weights.

8. The big data-driven intelligent system for dynamic hotel pricing and revenue optimization according to claim 7, characterized in that, The process of updating the preset sensitivity coefficient includes: When the order volatility index is less than the preset second volatility threshold At that time, the preset sensitivity coefficient corresponding to the score of the first demand index is triggered for iterative update, and the preset sensitivity coefficient after a single update is obtained. ,in, , To correspond to the preset sensitivity coefficient before a single update, The maximum allowed preset sensitivity coefficient. Adjust the step size for the preset coefficient; The order volatility index is obtained after each update until it meets a preset output condition. Then, the preset sensitivity coefficient of the score corresponding to the first demand index at that time is applied to the next preset monitoring period. The preset output condition is that the corresponding order volatility index is greater than or equal to a preset second volatility threshold. And less than or equal to the preset first fluctuation threshold .

9. A big data-driven intelligent method for dynamic hotel pricing and revenue optimization, characterized in that: Includes the following steps: Simultaneously acquire urban event data, regional meteorological data, and hotel traffic data within a preset monitoring period; Based on the city event data and regional meteorological data, event scores and meteorological scores are extracted respectively. Based on the meteorological scores and hotel traffic data, a traffic score is extracted. The scores are ranked to obtain the first demand index, the second demand index, and the third demand index. The dynamic demand index is obtained by combining the preset fusion weights of the corresponding order. The dynamic demand index and the first demand index are input into a preset pricing mapping model to output a pricing adjustment coefficient, which is then combined with the base price of the target room type to generate the corresponding preliminary price. Obtain the competitor pricing of the target room type within a preset monitoring period, and adjust the preliminary pricing by combining it with the preset sensitivity coefficient of the score corresponding to the first demand index to obtain its final pricing; Obtain the order volatility index of the target room type within a preset time window after the final pricing is executed, and update the preset fusion weight or preset sensitivity coefficient according to it until the order volatility index that meets the preset output conditions is obtained.

10. A computer storage medium storing computer-executable instructions, characterized in that, When the computer-executable instructions are executed, they realize the big data-driven intelligent system for dynamic hotel pricing and revenue optimization as described in any one of claims 1-8.

Images