Metro station surrounding bus stop reorganization optimization method based on passenger selection behavior

Abstract

The present application relates to the technical field of bus station reorganization optimization, and discloses a subway station surrounding bus station reorganization optimization method based on passenger selection behavior, which comprises the following steps: firstly, a decision model considering passenger path selection behavior is established to display a facility supply shaping passenger selection scheme model; the decision model and the scheme model are integrated to form a complete optimization model; and the optimization model is solved to calculate an optimal selection scheme. The present application compares the new model with existing methods and tests the sensitivity of the new model to budget and other factors. Finally, the present application provides practical suggestions, such as which stations should be upgraded first, to help cities make more intelligent investments in the field of public transportation.

Description

Technical Field

[0001] This invention relates to the field of bus stop reorganization and optimization technology, specifically a method for reorganizing and optimizing bus stops around subway stations based on passenger selection behavior. Background Technology

[0002] With rapid urbanization, subway systems have become the backbone of public transportation in large cities. Subways are renowned for their large capacity and reliability, but many passengers face the persistent "last mile" problem—the final leg of the journey from the subway station to their final destination is often too long to walk, bicycle availability is unreliable, and taxis are prohibitively expensive. When this connection fails, the overall efficiency of the subway system is diminished. The last mile problem may seem minor, but it impacts the entire transportation network. It extends travel time and may force people to revert to private car travel. Without addressing this issue, the substantial investments in subway systems risk being wasted. Therefore, improving last-mile connectivity is not merely a matter of convenience, but a key to building truly efficient and sustainable urban transportation systems.

[0003] Several solutions have been attempted to address the "last mile" problem, but the effectiveness of bike-sharing has been inconsistent, and ride-hailing services remain expensive. However, regular buses are often overlooked. They offer wide coverage, low fares, and seamlessly integrate with subway services. Better integration through methods such as coordinating timetables and unified fare systems can transform subway stations into well-connected hubs, rather than isolated nodes. However, this presents several practical challenges. A complete redesign of all bus stops around every subway station is prohibitively costly and impractical. Given limited resources, a crucial question arises: which stops should be prioritized for improvement? This is not a simple choice. Transportation planners must allocate limited funds where maximum benefit is achieved, while passengers choose routes based on convenience. This invention requires a clear approach to understanding the interplay between these two factors, and an effective way to identify which directions of improvement will have the greatest impact.

[0004] Research on public transport stop layout has been conducted at the macro, meso, and micro levels. Early studies focused on fundamental issues such as stop location selection and spacing. Later, research shifted to the coordination between public transport and rail transit. For example, designing routes that connect to subway lines rather than competing with them. Recent ideas include creating "micro-hubs" near subway stations to shorten walking distances and using smart technologies for dynamic scheduling. A common approach in transportation planning is the two-level optimization model. It helps analyze decision-making problems where planners make decisions at the upper level while traveler behaviors (such as route selection) are simulated at the lower level. This method has been applied to areas such as flexible public transport route design and multimodal network design. Improved algorithms make these models more practical. Nevertheless, current research still has a major limitation: most models fail to adequately consider passenger responses to improvements. They often treat passenger choices as fixed, ignoring the dynamic change that shorter transfer times may make buses more attractive. This one-sided view can lead to inefficient planning; for example, upgraded stops may not attract enough passengers. Furthermore, a comprehensive framework integrating investment, service improvement, passenger behavior, and system benefits is lacking.

[0005] Therefore, to address the above issues, a method for reorganizing and optimizing bus stops around subway stations based on passenger choice behavior is needed. Summary of the Invention

[0006] The purpose of this invention is to provide an optimization method for the reorganization of public transport stops around subway stations based on passenger choice behavior. This invention compares the new model with existing methods and tests its sensitivity to budget and other factors. Finally, this invention will provide practical recommendations, such as which stops should be prioritized for upgrades, to help cities make more informed investments in public transport.

[0007] This invention is implemented as follows: This invention provides a method for optimizing the reorganization of bus stops around subway stations based on passenger selection behavior, which is implemented according to the following steps: S1: First, establish a decision-making model that considers passenger route selection behavior, and then present a model of how facility supply shapes passenger choices. The decision-making model includes an upper-level model and a lower-level model: The upper-level model aims to minimize the average transfer distance. First, it determines the total transfer distance, as shown in the following formula:

[0008] The average transfer distance to the destination is calculated based on the total transfer distance, as shown in the following formula:

[0009] Since the denominator is a constant, the equation simplifies to minimizing the total transfer distance, as shown in the following formula:

[0010] Calculate passengers' transfer decisions ; The lower-level model includes a pattern selection model based on the Logit model, for a given upper-level decision Calculate the transfer distance for buses using the following formula:

[0011] Through utility function

[0012]

[0013] in, If the terms are independent and identically distributed Gumbel random terms, then the lower-level model is transformed as follows:

[0014] .

[0015] The sensitivity parameter θ captures passengers' true level of concern about transfer distance. A higher θ means people are more averse to longer walking distances, and even a small increase can alter their choice. Conversely, a lower θ indicates that passengers are less sensitive to distance, and their decisions are more random.

[0016] S2: Integrate the decision-making model and the solution model to form a complete optimization model; Integrating the upper and lower layers of the model, the resulting optimized model is as follows:

[0017] st ; in,

[0018] .

[0019] S3: Solve the optimization model to find the optimal choice. Follow these steps: S3.1: Quantify the system efficiency improvement that each site can achieve with a unit investment; first calculate the improvement for each site. i Calculate the value index as follows: ; S3.2: Sort all candidate sites in descending order according to the value index to form a priority queue. L ; S3.3: Perform a greedy selection; S3.4: Local search. Repeat each neighborhood operation, including addition, removal, and swap. If the operation improves the objective function value and meets the budget constraint, repeat until the stopping condition is met. If no improvement is found after K consecutive iterations during the process of repeating until the stopping condition is met, stop the search and impose resource constraints. Set a maximum number of iterations to avoid overcomputation. Terminate the process when the relative improvement of the objective function is lower than a preset value, based on the accuracy threshold.

[0020] Furthermore, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the above-described method for optimizing the reorganization of bus stops around subway stations based on passenger selection behavior.

[0021] Furthermore, the present invention provides a computer-storable medium storing a computer program, wherein when the program is executed, it sequentially executes any one of the above-described methods for reorganizing and optimizing bus stops around subway stations based on passenger selection behavior.

[0022] Compared with the prior art, the beneficial effects of the present invention are: 1. It overcomes the efficiency problems encountered when processing large-scale data. It constructs a complete decision-making process from parameter setting and scheme generation to result verification, helping to shift decision-making from reliance on experience to scientific methods. It provides a replicable framework for similar resource allocation problems.

[0023] 2. In the two-layer optimization model constructed in this invention, the goal of the upper-level planner is to minimize the total transfer distance of the entire network. Under strict budget constraints, binary decision variables are used to select subway stations for public transport service upgrades. This achieves a trade-off between passenger flow, cost, and expected benefits. Passengers, as individual decision-makers, minimize their travel costs by choosing buses or other modes of transportation given transfer conditions.

[0024] 3. This invention seeks to reduce the average transfer distance between subway and bus services without exceeding the available budget. The upper-level objective of this invention is to minimize the weighted sum of the expected transfer distances for all passengers across the entire network. Passenger flow and corresponding transfer distances at each station are included in the calculation. A key constraint is that the total cost of the selected stations must be kept within the budget, reflecting the scarcity of resources in reality and forcing planners to carefully determine priorities. The objective and constraints together define a classic resource-constrained combination, and this invention realizes how to allocate limited resources to maximize the overall system efficiency. Attached Figure Description

[0025] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0026] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to describe selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Please see Figure 1 This invention provides a method for optimizing the reorganization of bus stops around subway stations based on passenger selection behavior, which is specifically implemented according to the following steps: S1: First, establish a decision-making model that considers passenger route selection behavior, and then present a model of how facility supply shapes passenger choices. The decision-making model includes an upper-level model and a lower-level model: The upper-level model aims to minimize the average transfer distance. First, it determines the total transfer distance, as shown in the following formula:

[0029] in, Representatives gathered at the subway station. Represents a set of transportation modes, among which ; For the site The number of passengers; and Site Bus transfer distances before and after renovation; Site Transfer distances between other modes of transportation; The average transfer distance to the destination is calculated based on the total transfer distance, as shown in the following formula:

[0030] Since the denominator is a constant, the equation simplifies to minimizing the total transfer distance, as shown in the following formula:

[0031] in, For the site The cost of renovating bus stops; This is the total investment budget; Calculate passengers' transfer decisions ; The lower-level model includes a pattern selection model based on the Logit model, for a given upper-level decision Calculate the transfer distance for buses using the following formula:

[0032] Through utility function

[0033]

[0034] in, For the sensitivity parameter in the Logit model, If the terms are independent and identically distributed Gumbel random terms, then the lower-level model is transformed as follows:

[0035] .

[0036] The sensitivity parameter θ captures passengers' true level of concern about transfer distance. A higher θ means people are more averse to longer walking distances, and even a small increase can alter their choice. Conversely, a lower θ indicates that passengers are less sensitive to distance, and their decisions are more random.

[0037] S2: Integrate the decision-making model and the solution model to form a complete optimization model; Integrating the upper and lower layers of the model, the resulting optimized model is as follows:

[0038] st ; in,

[0039] .

[0040] Among them, the upper-level binary variables Indicates whether or not on the site Bus stop renovations are underway. (Lower-level continuous variables) and They represent the sites The proportion of passengers choosing buses versus other modes of transportation.

[0041] S3: Solve the optimization model to find the optimal selection scheme. In this embodiment, firstly, all candidate sites are sorted in descending order according to the value index to form a priority queue. Next, an empty solution set and remaining budget are initialized. Then, starting from the top of the queue, each site is evaluated one by one. If the renovation cost of a site does not exceed the remaining budget, it is added to the solution set, and the budget is reduced accordingly. This selection process is iterated until all candidate sites have been considered, or the remaining budget cannot cover any remaining sites. Finally, the set of selected sites is output as the initial feasible solution. The main advantages of this invention are its simplicity and efficiency. The time complexity is only [missing information]. .

[0042] In the Logit-based model, the sensitivity parameter θ captures passengers' true level of concern about transfer distance. A higher θ means people are more averse to longer walking distances, and even a small increase can alter their choice. Conversely, a lower θ indicates that passengers are less sensitive to distance, and their decisions are more random.

[0043] In this embodiment, non-public transportation modes such as walking, cycling, and taxis are grouped together. First, this maintains the simplicity of the model, avoiding the complexity of nested decision structures. Second, it reflects people's decision-making process: in last-mile travel, many people first choose between taking public transportation or not, and only then select from other options. Grouping them into one category simply follows this natural way of thinking, specifically implemented through the following steps: S3.1: Quantify the system efficiency improvement that each site can achieve with a unit investment; first calculate the improvement for each site. i Calculate the value index as follows: ; S3.2: Sort all candidate sites in descending order according to the value index to form a priority queue. L ; S3.3: Perform a greedy selection; In this embodiment, the algorithm calculates the value index, sorts the sites, and performs a greedy selection. Calculating the value index requires traversing all n sites, with a constant time consumption for each site, thus this step is O(n). Quicksort is used for sorting, with an average time complexity of O(n log n). The greedy selection step requires scanning the sorted list again, also taking O(n). Overall, the total time complexity is determined by the highest-order term, which is O(n log n). This demonstrates the algorithm's good scalability; even for large-scale problems, it can still produce solutions in a reasonable amount of time, making it suitable for practical applications.

[0044] S3.4: Local search. Repeat each neighborhood operation, including addition, removal, and swap. If the operation improves the objective function value and meets the budget constraint, repeat until the stopping condition is met. If no improvement is found after K consecutive iterations during the process of repeating until the stopping condition is met, stop the search and impose resource constraints. Set a maximum number of iterations to avoid overcomputation. Terminate the process when the relative improvement of the objective function is lower than a preset value, based on the accuracy threshold.

[0045] The specific operations for the local search described above are as follows: Input: Site collection ,Budget ,cost Passenger volume ,distance ,parameter ; Output: The set of selected sites ; 1. For each site Computational Value Index:

[0046] 2. Sort sites in descending order of value index → ​​List

[0047] 3. Greedy Choice: remaining_budget =

[0048] for in : if ≤ remaining_budget:

[0049] remaining_budget =

[0050] Local search: Repeat the following steps until the stopping condition is met: For each neighborhood operation (add, remove, swap): If the operation improves the objective function value and satisfies the budget constraint: Apply this operation 5. Return

[0051] In the above embodiments, the maximum number of iterations and the accuracy threshold are determined based on empirical data from a long-term public transport history database.

[0052] In this embodiment, the current method is calculated using specific instance data: Based on actual data from a major city's rail transit system, a case study was constructed comprising 50 subway stations. The parameters of each station varied significantly: daily passenger flow ranged widely, from 68,000 to 163,000, with an average of 112,600 passengers per day. The renovation costs, based on actual project data, ranged from 260,000 to 960,000 RMB. Before the renovation, bus transfer distances also varied significantly, ranging from 140 meters to 450 meters, with an average of 287.3 meters. After the renovation, following the principle of "zero-distance transfer," transfer distances were set between 28 and 90 meters. This setting reflects the diverse characteristics of stations located in different areas and with different functions, providing a realistic data foundation for model validation.

[0053] As shown in Table 1, in the baseline scenario without any modifications, the average transfer distance is 187.3 meters, and the public transport modal share is only 42.1%. Analysis reveals that at eight stations with transfer distances exceeding 350 meters, the public transport modal share drops below 30%; while at 15 stations with transfer distances less than 200 meters, the modal share exceeds 50%. This clearly highlights the urgency of improving long-distance transfer stations.

[0054] Table 1 Key Indicators for Baseline Scenario

[0055] As shown in Table 2, under a budget constraint of 20 million yuan, the optimization algorithm selected 14 sites for renovation, with a total cost of 19.82 million yuan. The selected sites clearly possess "high-value" characteristics: an average daily passenger flow of 142,000, an average improvement potential of 342 meters, and an average value index of 1.86. These sites are mostly located in core business districts, transportation hubs, and large residential areas, reflecting a balance between demand orientation and network synergy.

[0056] Table 2 Selected Site Features in the Optimization Scheme

[0057] Table 3 Performance Evaluation of Optimization Scheme

[0058] As shown in Table 4, the investment benefit analysis indicates that the scheme has good cost-effectiveness: each unit of investment (RMB 10,000) can reduce the transfer distance by 154.6 meters. The 14 upgraded stations serve 1.58 million passengers daily, accounting for 28.1% of the total system passenger flow. It is worth noting that station No. 8, with the highest cost (RMB 980,000), was still selected because of its great improvement potential (360 meters) and high passenger flow (163,000 passengers / day). The optimization results reveal a reasonable selection process. For example, although station No. 1 has 125,000 passengers daily, it was not selected due to its high cost (RMB 850,000) and low value index (0.89). In contrast, station No. 31, with less passenger flow (142,000 passengers / day), was prioritized because of its excellent cost-effectiveness (value index 2.1). This shows that the algorithm does not simply select the station with the highest passenger flow, but rather balances improvement potential, cost-effectiveness, and network effects.

[0059] Table 4 Investment Benefit Analysis

[0060] As shown in Table 5, budget sensitivity analysis indicates that the return on investment decreases with increasing budget. When the budget increases from 15 million yuan to 20 million yuan, the improvement rate of transfer distance jumps from 24.0% to 29.2%. However, when the budget further increases to 25 million yuan, the improvement rate only reaches 32.3%. This suggests that a budget of 20 million yuan already covers most high-value stations, and further investment requires more careful consideration. Parameter sensitivity analysis shows that the distance sensitivity parameter... It has a significant impact. When When the value increased from 0.008 to 0.015, the average transfer distance decreased from 138.2 meters to 121.4 meters, meaning that the more sensitive passengers are to distance, the more significant the effect of the improvement. This finding emphasizes that investment should be prioritized in areas where passengers are more sensitive to walking distance. The decreasing trend is evident in Table 5: after exceeding 20 million yuan, the improvement brought by each additional 5 million yuan gradually decreases (e.g., from 29.2% to 32.3%). A clear saturation point appears around 25 million yuan; beyond this point, the improvement brought by each million yuan of investment is less than 1%. This indicates that for this network, the optimal budget range is between 20 and 25 million yuan, and investments exceeding this range should be redirected to other transportation projects.

[0061] Table 5 Budget Sensitivity Analysis

[0062] As shown in Table 6, based on the optimization results, this invention proposes a four-tier investment priority strategy: the first tier includes stations with a value index higher than 2.0 (immediate investment); the second tier covers stations with a value index of 1.5-2.0 (priority allocation); the third tier includes stations with a value index of 1.0-1.5 (considered on a case-by-case basis); and the fourth tier includes stations with a value index lower than 1.0 (investment postponed). This strategy provides a clear decision-making framework for urban public transportation investment. This optimization model provides a scientific basis for upgrading public transportation stations in large cities. By accurately identifying high-value stations, system benefits can be maximized within a limited budget. If implemented, this scheme is expected to reduce transfer distances by 11.18 billion meters annually, significantly improving the attractiveness of public transportation and supporting greener urban travel. The flexibility of this model also allows it to meet the specific needs of different cities.

[0063] Table 6 Recommended Site Investment Priorities

[0064] Using the example data above, a complete decision-making process was constructed, from parameter setting and scheme generation to result verification, which helps to promote the transformation of decision-making from relying on experience to scientific methods. These contributions provide a replicable framework for similar resource allocation problems.

[0065] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations will be apparent to those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for optimizing the reorganization of bus stops around subway stations based on passenger choice behavior, characterized by: Follow these steps: S1: First, establish a decision-making model that considers passenger route selection behavior, and then present a model of how facility supply shapes passenger choices. S2: Integrate the decision-making model and the solution model to form a complete optimization model; S3: Solve the optimization model to find the optimal choice.

2. The method for reorganizing and optimizing bus stops around subway stations based on passenger selection behavior according to claim 1, characterized in that: The decision-making model includes an upper-level model and a lower-level model: The upper-level model aims to minimize the average transfer distance. First, it determines the total transfer distance, as shown in the following formula: ； The average transfer distance to the destination is calculated based on the total transfer distance, as shown in the following formula: ； Since the denominator is a constant, the equation simplifies to minimizing the total transfer distance, as shown in the following formula: ； Calculate passengers' transfer decisions ; The lower-level model includes a pattern selection model based on the Logit model, for a given upper-level decision Calculate the transfer distance for buses using the following formula: ； Through utility function ； ； in, If the terms are independent and identically distributed Gumbel random terms, then the lower-level model is transformed as follows: ； 。 3. The method for reorganizing and optimizing bus stops around subway stations based on passenger selection behavior according to claim 1, characterized in that: In step S2, the upper and lower layer models are integrated to form the following optimized model: ； s.t. ; in, ; 。 4. The method for reorganizing and optimizing bus stops around subway stations based on passenger selection behavior according to claim 1, characterized in that: In step S3, the following steps are specifically performed: S3.1: Quantify the system efficiency improvement that each site can achieve with a unit investment; first calculate the improvement for each site. i Calculate the value index as follows: ； S3.2: Sort all candidate sites in descending order according to the value index to form a priority queue. L ; S3.3: Perform a greedy selection; S3.4: Local search, repeat each neighborhood operation, including addition, removal, and swap. If the operation improves the objective function value and satisfies the budget constraint, repeat until the stopping condition is met.

5. The method for reorganizing and optimizing bus stops around subway stations based on passenger selection behavior according to claim 4, characterized in that: If no improvement is found after K consecutive iterations during the process of repeating until the stopping condition is met, the search is stopped and resource constraints are imposed. A maximum number of iterations is set to avoid overcomputation. Based on an accuracy threshold, the process terminates when the relative improvement of the objective function is lower than a preset value.

6. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for reorganizing and optimizing bus stops around subway stations based on passenger selection behavior as described in any one of claims 1-5.

7. A computer-storable medium storing a computer program therein, characterized in that: When the program is executed, it sequentially executes any one of the above claims 1-5, the method for reorganizing and optimizing bus stops around subway stations based on passenger selection behavior.

Images