A collaborative optimization method for the layout of electric bicycle facilities based on multi-dimensional features

Abstract

This invention proposes a collaborative optimization method for electric bicycle facility layout based on multi-dimensional features, relating to the field of urban transportation planning technology. The method includes: obtaining the comprehensive demand potential index of candidate locations and decomposing it into parking facility demand, shared facility demand, and charging facility demand; constructing a collaborative efficiency gain function for each candidate location; constructing a service interference loss function for each candidate location; constructing a facility configuration efficiency index for each candidate location; establishing a multi-objective programming model; solving the multi-objective programming model using a multi-objective genetic algorithm to obtain a Pareto optimal solution set; identifying suitable regional characteristics for composite and single configurations based on the facility configuration efficiency index values ​​from the Pareto optimal solution set, and outputting the facility type configuration and deployment scale for each candidate location. This invention solves the technical problem of the lack of quantitative basis for composite site configuration in existing technologies.

Description

Technical Field

[0001] This invention relates to the field of urban transportation planning technology, and in particular to a collaborative optimization method for the layout of electric bicycle facilities based on multi-dimensional features. Background Technology

[0002] With the rapid growth of urban electric bicycle ownership, the scientific layout of infrastructure such as parking spots, sharing stations, and charging piles has become a crucial issue in urban traffic management. Existing electric bicycle facility layout methods mainly fall into two categories: one involves planning various facilities independently, with parking spots, sharing stations, and charging piles scattered across different plots of land. This approach leads to redundant land use and infrastructure construction, resulting in low space utilization efficiency. The other approach simply concentrates various facilities into composite sites, assuming that concentration is always better than dispersion. However, this ignores the potential service interference between different facility types within composite sites. When private vehicle parking, shared vehicle rental and charging operations compete spatially, it leads to a decline in user experience and reduced operational efficiency. The fundamental flaw of these two models lies in the lack of a quantitative trade-off mechanism for the positive and negative effects of coordinated facility configuration.

[0003] Chinese patent CN111582552A discloses a method for allocating shared bicycle parking spots based on a multi-objective genetic algorithm. This patent collects user request data and parking spot location information, establishes a bi-objective optimization model with the objectives of minimizing distance and density costs, uses the NSGA-II algorithm to match parking spots for vehicles in the request queue, and improves the algorithm's convergence by incorporating a regression algorithm to enhance the mutation operator. While this patent shows some effectiveness in processing multiple user requests in a short time, the method only dynamically allocates parking spots for a single type of shared bicycle and does not address the collaborative layout of multiple types of facilities. It fails to solve the problems of how to spatially combine and configure multiple types of facilities and how to determine the scale of each facility. Summary of the Invention

[0004] In view of this, the present invention provides a method for collaborative optimization of electric bicycle facility layout based on multi-dimensional features. By integrating multi-dimensional features of the built environment to identify candidate locations, a collaborative efficiency gain function is established to quantify the positive collaborative effect of composite sites, and a service interference loss function is established to quantify the negative interference effect of composite sites. A facility configuration efficiency index that balances the two is constructed as the optimization target. This enables a scientific layout of deploying composite sites in high-demand and high-collaboration areas and deploying single facilities in areas with dispersed demand or interference sensitivity, thus solving the technical problem of lack of quantitative basis for the configuration of composite sites in the prior art.

[0005] The technical solution of this invention is implemented as follows:

[0006] This invention provides a collaborative optimization method for the layout of electric bicycle facilities based on multi-dimensional features, comprising:

[0007] S1. Based on road network density characteristics, point of interest distribution characteristics, public transport station accessibility characteristics and population density characteristics, obtain the comprehensive demand potential index of candidate locations and decompose it into parking facility demand, shared facility demand and charging facility demand.

[0008] S2. Based on the degree of resource reuse, land intensive use and scale synergy among different facility types, construct a synergy efficiency gain function for each candidate location;

[0009] S3. Based on the degree of service interference, capacity pressure level and coupling degree of parking facility demand, shared facility demand and charging facility demand among different facility types, construct a service interference loss function for each candidate location.

[0010] S4. Construct the facility configuration efficiency index for each candidate location based on the collaborative efficiency gain function and the service interference loss function;

[0011] S5. With the objectives of maximizing the overall system configuration efficiency, maximizing the demand-weighted coverage rate, and minimizing the total construction cost, and with constraints of scale, spatial capacity, coverage, budget, and spacing, a multi-objective programming model is established; wherein the overall system configuration efficiency is the sum of the facility configuration efficiency indices of each candidate location, and the weight of the demand-weighted coverage rate is the comprehensive demand potential index.

[0012] S6. Solve the multi-objective programming model using a multi-objective genetic algorithm to obtain the Pareto optimal solution set;

[0013] S7. Identify suitable regional characteristics of composite and single configurations from the Pareto optimal solution set based on the facility configuration efficiency index value, and output the facility type configuration and deployment scale of each candidate location.

[0014] Preferably, the comprehensive demand potential index of candidate locations obtained in step S1 based on road network density characteristics, point-of-interest distribution characteristics, public transport stop accessibility characteristics, and population density characteristics includes:

[0015] The study area was divided into square grid cells. For each grid cell, the road network density features, point of interest distribution features, public transport station accessibility features, and population density features were extracted. After normalizing each feature, a weighted summation method was used to calculate the comprehensive demand potential index.

[0016] The road network density features include road density, non-motorized vehicle lane coverage, and road network connectivity index; the point of interest distribution features include residential area density, commercial area density, office area density, and educational facility density; the public transport stop accessibility features include public transport stop density, public transport route coverage, and average distance between public transport stops; and the population density features include resident population density and employed population density.

[0017] Preferably, in step S1, the comprehensive demand potential index is decomposed into parking facility demand, shared facility demand, and charging facility demand, including:

[0018] For each candidate location, the comprehensive demand potential index of each grid cell within the service range is weighted and summed using the distance decay function to obtain the total demand within the service range of that candidate location.

[0019] Based on the density of residential areas, commercial areas, office areas, and educational facilities in the grid unit where the candidate location is located, a demand decomposition rule is established and the total demand is decomposed into three types of facility demand. Among them, the density of residential areas dominates the demand for parking facilities, the density of commercial areas dominates the demand for shared facilities, and the density of office areas and educational facilities jointly affect the demand for charging facilities.

[0020] Preferably, in step S2, the synergistic performance gain function is:

[0021] ;

[0022] Where j is the candidate position index; k and K is the facility type index; K is the set of facility types, including parking spots, shared stations, and charging stations. The decision variable is defined as whether facility type k is deployed at candidate location j. For facility type k and The basic synergy coefficient between them reflects the degree of resource reuse; As a factor for intensive land use; The scale synergy factor is calculated as follows:

[0023] ;

[0024] In the formula, The scale of facility type k deployed for candidate location j; This is a reference size for facility type k; This represents the total number of facility types.

[0025] Preferably, in step S3, the service interference loss function is:

[0026] ;

[0027] Where j is the candidate position index; k and K is the facility type index; K is the set of facility types; To define decision variables; For facility type k and The basic interference coefficient between them reflects the degree of service interference; Capacity pressure factor; The demand coupling factor is calculated as follows:

[0028] ;

[0029] In the formula, The demand for facility type k within the service area of ​​candidate location j. For the reference size of facility type k, The scale of facility type k deployed for candidate location j; The service capacity per unit size for facility type k; This is the demand pressure coefficient; To prevent extremely small positive numbers with a denominator of zero.

[0030] Preferably, step S4 includes:

[0031] The synergy performance gain function is converted into a synergy performance gain index, and its expression is as follows:

[0032] ;

[0033] Where j is the candidate location index; k is the facility type index; and K is the set of facility types, including parking spots, shared stations, and charging piles. The synergistic performance gain function; To define decision variables; The scale of facility type k deployed for candidate location j; The basic resource allocation for facility type k; This is a reference size for facility type k; This refers to the amount of resources allocated per unit of scale. To improve the collaborative savings rate;

[0034] The service interference loss function is converted into an interference effectiveness loss metric, the expression of which is:

[0035] ;

[0036] in, For service interference loss function; The service capacity per unit size for facility type k; Service capacity factor per unit; To compensate for the loss rate;

[0037] The facility configuration efficiency index is defined as follows:

[0038] ;

[0039] in, As a synergistic performance gain indicator, This is an indicator of interference effectiveness loss.

[0040] Preferably, in step S5, the objective function of the multi-objective programming model is:

[0041] ;

[0042] ;

[0043] in, L represents the overall configuration efficiency; L is the set of candidate positions; j is the index of the candidate position. The facility configuration efficiency index for candidate location j; The coverage is a demand-weighted ratio; N is the total number of grid cells; i is the grid cell index. The comprehensive demand potential index for grid cell i; This is an auxiliary coverage variable, indicating whether grid cell i is covered.

[0044] Preferably, the constraints of the multi-objective programming model include:

[0045] Size constraints Limit the deployment scale of each facility type;

[0046] Space capacity constraints Limit the space occupancy rate of composite sites;

[0047] Coverage constraints Define the sufficiency of grid coverage;

[0048] Budget constraints The total resource investment in the construction of control facilities shall not exceed the budget limit B;

[0049] Spacing constraints Avoid excessive concentration of similar facilities;

[0050] Where K is the set of facility types, and k is the facility type index. and Let these be the minimum and maximum sizes of facility type k, respectively. Let j be the actual area occupied by the composite site at candidate location j. For usable area, To maximize the use of space, For candidate position index, Let j be the distance from the candidate position j to the center of grid cell i. Let B be the maximum service radius of facility type k, and B be the total budget ceiling. The basic resource allocation for facility type k. To define decision variables; Let k be the unit size resource allocation for facility type. The scale of facility type k deployed for candidate location j. The site development cost for candidate location j. For indicator functions, For candidate positions j and The distance between them The minimum allowable spacing for facility type k.

[0051] Preferably, in step S6, the multi-objective genetic algorithm uses the NSGA-II algorithm; chromosomes are encoded with integers to represent the scale configuration of each facility type at each candidate location, with a gene position value of 0 indicating no facility of that type is deployed, and a positive integer indicating the deployment of facilities of the corresponding scale; the optimization strategy combines tournament selection, two-point crossover, and uniform mutation, and fitness is evaluated using fast non-dominated sorting and crowding distance indicators; the algorithm prioritizes individuals with low non-dominated levels and large crowding distances; the optimization stopping condition is that the rate of change of the hypervolume index of the non-dominated solution set for M consecutive generations is less than 1%. .

[0052] Preferably, step S7 includes:

[0053] The facility configuration efficiency index values ​​for each candidate location are statistically analyzed from the Pareto optimal solution set.

[0054] Candidate locations with positive facility configuration efficiency index values ​​are selected, and their comprehensive demand potential index, available site area, point of interest functional mixing degree, and public transport accessibility characteristics are extracted.

[0055] Based on the extracted features, cluster analysis is used to determine the feature thresholds of suitable areas for composite configuration, and to identify high-demand density areas, areas around public transport hubs, and mixed-use areas.

[0056] Based on the facility configuration efficiency index value and the identified feature threshold, the facility type configuration and deployment scale of each candidate location are output. Candidate locations with a positive facility configuration efficiency index value and meeting the feature threshold are deployed with composite sites, while other candidate locations are deployed with single-type facilities.

[0057] The present invention has the following advantages over the prior art:

[0058] This invention quantifies the land-saving, infrastructure-sharing, and scale synergy effects of composite sites by constructing a collaborative efficiency gain function, and simultaneously constructs a service interference loss function to quantify the service efficiency decline caused by spatial competition, capacity pressure, and demand overload. Based on the difference between these two functions, a facility configuration efficiency index is defined to achieve a quantitative trade-off between composite and single configurations. This allows for the deployment of composite sites in locations with high demand and strong synergy to improve resource utilization efficiency, while deploying single facilities in locations with dispersed demand or interference sensitivity to ensure service quality. The multi-objective optimization model maximizes system configuration efficiency while considering demand coverage and construction cost constraints. The output Pareto optimal solution set provides differentiated decision-making schemes for different budget and preference scenarios. Compared to extreme strategies of all composite or all single facilities, this method has significant advantages in reducing service interference losses, improving collaborative efficiency gains, and enhancing overall system configuration efficiency. Attached Figure Description

[0059] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0060] Figure 1 This is a flowchart of the method of the present invention;

[0061] Figure 2 This is a diagram illustrating the technical implementation of the present invention;

[0062] Figure 3 This is a flowchart of the algorithm of the present invention. Detailed Implementation

[0063] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0064] like Figure 1As shown, this invention provides a method for collaborative optimization of electric bicycle facility layout based on multi-dimensional features, including: S1, obtaining a comprehensive demand potential index for candidate locations based on road network density features, point of interest distribution features, public transport stop accessibility features, and population density features, and decomposing it into parking facility demand, shared facility demand, and charging facility demand; S2, constructing a collaborative efficiency gain function for each candidate location based on the degree of resource reuse, land intensive use, and scale synergy among different facility types; S3, constructing a service interference loss function for each candidate location based on the degree of service interference, capacity pressure level, and coupling degree of the parking facility demand, shared facility demand, and charging facility demand among different facility types; S4, based on the collaborative efficiency gain function and the service interference... S5. A multi-objective programming model is established with the objectives of maximizing the overall system configuration efficiency, maximizing the demand-weighted coverage, and minimizing the total construction cost, and with constraints of scale, spatial capacity, coverage, budget, and spacing. The overall system configuration efficiency is the sum of the facility configuration efficiency indices of each candidate location, and the weight of the demand-weighted coverage is the comprehensive demand potential index. S6. A multi-objective genetic algorithm is used to solve the multi-objective programming model to obtain the Pareto optimal solution set. S7. Based on the facility configuration efficiency index values, suitable regional characteristics of composite and single configurations are identified from the Pareto optimal solution set, and the facility type configuration and deployment scale of each candidate location are output.

[0065] like Figure 2 As shown, the technical implementation process of this invention is as follows: First, through step S1, multi-dimensional features are extracted from urban built environment data to quantify the demand potential of each candidate location and decompose it into differentiated demands for three types of facilities. These demand indicators serve as known input data for subsequent function construction. Second, through steps S2 to S4, an evaluation system is established, constructing a system based on decision variables. and The independent variables are the synergy efficiency gain function, the service interference loss function, and the facility configuration efficiency index. Model parameters within these functions, such as synergy coefficients and interference coefficients, were calibrated using expert surveys and historical data and are determined to be constants before optimization. These functions define the evaluation rules for the deployment scheme, but the specific deployment scheme has not yet been determined at this stage. Then, step S5 establishes a multi-objective optimization model centered on the facility configuration efficiency index, with decision variables... and The unknowns to be solved are: In step S6, a multi-objective genetic algorithm is used iteratively to solve the problem. In each generation of the algorithm, each individual represents a specific set of decision variable values, i.e., a complete hypothetical layout scheme. The decision variable values ​​of this scheme are substituted into the function formulas defined in steps S2 to S4 for calculation, successively obtaining the collaborative efficiency gain, service interference loss, and configuration efficiency values ​​of each candidate location. These are then substituted into the objective function formula in step S5 to calculate the three objective function values: the overall system configuration efficiency, demand-weighted coverage, and total construction cost of the scheme. Based on the objective function values, the fitness of the individual is evaluated through non-dominated sorting and congestion distance calculation. After selection, crossover, and mutation operations, a new generation of population is generated, and the above calculation process is repeated. After multiple generations of iteration, the algorithm outputs the Pareto optimal solution set. Finally, step S7 analyzes the optimal solution set, identifying suitable areas for composite and single configurations based on the facility configuration efficiency index values ​​of each candidate location in the optimal solution, and outputting recommended facility type configuration and deployment scale schemes.

[0066] In one embodiment of the present invention, step S1 includes:

[0067] The study area was divided into square grid units with a side length of 500 meters, and the grid size was adjusted between 300 meters and 1000 meters according to the city size and data accuracy.

[0068] For each grid cell i, four types of built environment features are extracted to assess its potential demand for electric bicycle facilities:

[0069] The first category is road network density characteristics, including road density. Non-motorized vehicle lane coverage and road network connectivity index Road density is calculated as the ratio of the total length of roads within a grid to the grid area. Non-motorized vehicle lane coverage is calculated as the proportion of roads with non-motorized vehicle lanes to the total road length. The road network connectivity index is calculated as the ratio of the number of road intersections to the number of road segments; a higher index indicates a denser road network and better accessibility.

[0070] The second category is the distribution characteristics of points of interest, including residential area density. Density of commercial areas Office density and density of educational facilities The kernel density estimation method is used to calculate the density values ​​of corresponding interest points within the grid and an 800-meter radius. Areas with high residential density have strong parking demand, areas with high commercial density have strong sharing demand, and areas with high density of office and educational facilities have strong commuting charging demand.

[0071] The third category is the accessibility characteristics of public transport stops, including the density of public transport stops. Public transport route coverage Average distance between bus stops Bus stop density is calculated as the ratio of the number of bus stops within a grid to the grid area. Bus route coverage is calculated as the number of bus routes within a grid. Average bus stop spacing is calculated as the average distance between adjacent bus stops. Bus characteristics reflect the potential demand for electric bicycles as a public transport connection.

[0072] The fourth category is population density characteristics, including resident population density. and employment population density ;

[0073] After normalizing each feature, a weighted summation method is used to calculate the comprehensive demand potential index of each grid cell i. Defined as:

[0074] ;

[0075] in arrive The weighting coefficients for each feature are given, and the sum of all weights equals 1; Comprehensive Demand Potential Index The value reflects the demand intensity of electric bicycle facilities in the grid unit, and ranges from 0 to 1;

[0076] The method for determining the weight coefficients is as follows: First, a correlation analysis is performed on each feature to remove redundant features with a correlation coefficient greater than 0.8. Then, the analytic hierarchy process is used to invite 5 to 7 experts in urban transportation planning to compare and score the importance of each feature pairwise. After passing the consistency test, the feature weights are calculated. In practical applications, the weights can be adjusted according to the characteristics of the city. For example, cities with well-developed public transportation systems should increase the weight of public transportation features, while newly developed areas should increase the weight of road network features.

[0077] Based on the demand potential index and on-site surveys, a set of candidate locations was determined. The selection of candidate locations follows three principles: First, priority is given to locations with high demand potential. The first requirement is that the grid area is above the regional average. Secondly, the candidate location should possess basic land conditions, including available vacant land, roadside space, or open space near buildings, and a usable area. Not less than the minimum area threshold , The possible value is 100m. 2 Third, the candidate location is not located on the main traffic artery's motor vehicle lane or in a legally prohibited parking area;

[0078] For each candidate location j, the comprehensive demand potential index of each grid cell within the service range is weighted and summed using the distance decay function to obtain the total demand within the service range of that candidate location, defined as:

[0079] ;

[0080] in The set of grids covered by the service buffer for candidate location j, with a buffer radius of 800m. Let j be the Euclidean distance from candidate position j to the center of grid i, and the weight function be:

[0081] ;

[0082] in The distance attenuation parameter is set to 400m. This weighting function reflects the principle that demand decreases as distance increases.

[0083] The set of facility types is denoted as , where 1 represents a parking spot, 2 represents a shared station, and 3 represents a charging station;

[0084] Based on the density of residential areas, commercial areas, office areas, and educational facilities in the grid cells where the candidate locations are located, establish demand decomposition rules and calculate the total demand. The demand for facilities can be broken down into three categories: residential area density dominates the demand for parking facilities, commercial area density dominates the demand for shared facilities, and office area density and educational facility density jointly affect the demand for charging facilities.

[0085] Parking demand is defined as:

[0086] ;

[0087] The sharing requirement is defined as:

[0088] ;

[0089] Charging demand is defined as:

[0090] ;

[0091] in To prevent the denominator from being zero, the value is 0.01, which is the smallest positive number. , , , These represent the normalized residential area density, commercial area density, office area density, and educational facility density of the grid where candidate location j is located, respectively.

[0092] The decomposition coefficients reflect the demand preferences of different areas for different facility types. Residential areas contribute 1.0 to parking demand, representing a full contribution; office areas contribute 0.5 each to parking and charging, reflecting a diversion effect; commercial areas contribute 1.0 to sharing demand; and educational facilities contribute 0.3 and 0.7 to sharing and charging respectively, reflecting student commuting characteristics. The demand decomposition formula ensures that the sum of the three types of demand equals the total demand, conforming to the principle of demand conservation.

[0093] In one embodiment of the present invention, step S2 includes:

[0094] When multiple types of facilities are deployed simultaneously at candidate location j, a synergistic efficiency gain is generated. This gain stems from land resource sharing to avoid duplication of land use, infrastructure sharing to reduce construction costs, and management resource sharing to reduce operating costs, reflecting the positive effect of synergistic facility deployment. The synergistic efficiency gain function for candidate location j is defined as follows:

[0095] ;

[0096] Where j is the candidate position index; k and K is the facility type index; K is the set of facility types, including parking spots, shared stations, and charging stations. Let be the decision variable, representing whether facility type k should be deployed at candidate location j. This indicates that a facility of type k is deployed at candidate location j. This indicates that no deployment will be made;

[0097] For facility type k and The basic coordination coefficient between parking spots and charging piles reflects the degree of resource reuse in combinations of different facility types. The specific value can be determined according to the following rules: Coordination coefficient between parking spots and charging piles. The highest value reflects the complementary functions of both parking spots and shared stations, which both serve private electric bicycles and share land and power infrastructure. Centered, reflecting the partial sharing of parking space occupied by shared vehicles upon return, and the availability of shared management manpower, as well as the coordination coefficient between shared stations and charging piles. The relatively low value reflects the fact that while shared vehicles occupy charging stations, there are fundamental differences in the management logic despite the shared infrastructure; for example, the possible values ​​are... These coefficient values ​​can be calibrated and adjusted based on actual operational data in different cities.

[0098] The land-intensive use factor, which characterizes the proportion of land area saved by composite sites relative to their dispersed deployment, is defined as:

[0099] ;

[0100] in This refers to the actual area occupied by the composite site. This refers to the total area occupied when facilities of the same size are distributed in a dispersed manner. To prevent extremely small positive numbers with a denominator of zero;

[0101] The total area occupied when deployed in a dispersed manner is calculated as follows:

[0102] ;

[0103] in The scale of facility type k deployed for candidate location j; the scale of parking spot is the number of parking spaces, the scale of shared station is the number of vehicles, and the scale of charging pile is the number of charging piles. Let k be the standard area occupied per unit size of facility type k, for example, This indicates that the standard area of ​​each parking space at the parking lot is 6 square meters. 2 Includes parking spaces and access routes; each shared parking space has a standard area of ​​2.5 square meters. 2 Reflecting the close arrangement of shared vehicles, each charging station has a standard area of ​​4 square meters. 2 Includes charging equipment and operating space;

[0104] The maximum shareable area is calculated as follows:

[0105] ;

[0106] in For facility type k and The spatial superposition coefficient between the parking point and the charging pile is, in this embodiment, the superposition coefficient between the parking point and the charging pile. Setting it to the highest value reflects the maximum degree of spatial overlap that charging piles can achieve within parking spaces, representing the superposition coefficient of shared stations and charging piles. Centered, reflecting the overlap of space occupied by shared vehicles during charging, and the superposition coefficient of parking spots and shared stations. The lower percentage reflects that shared vehicles can utilize some parking spaces, but the functional differences between the two result in limited overlap; for example... ;

[0107] The actual area occupied by the composite site is:

[0108] ;

[0109] in This is the site space utilization efficiency coefficient, which can be taken as 0.25. Given the available area of ​​candidate location j, this formula shows that the area that can be saved by spatial stacking of composite sites is subject to the upper limit of the shareable area. and site layout constraints The dual restrictions;

[0110] Land intensive use factor The value ranges from 0 to 1, when only a single facility is deployed. ,therefore No shared gain, when multiple facilities are deployed ,therefore And it increases with the number of shared facilities;

[0111] The scale synergy factor is calculated as follows:

[0112] ;

[0113] in This is a reference size for facility type k, and its specific value can be set to... The reference size for parking lots is 50 parking spaces, for shared parking stations it's 30 vehicles, and for charging stations it's 20 charging piles. These reference sizes represent the minimum scale required for economical operation of the facilities. The total number of facility types is equal to 3. The scale synergy factor adopts the geometric mean form with an added power-law amplification effect. The synergy gain is maximized when the scale of various facilities grows in a balanced manner. When the scale of a certain type of facility is much smaller than the reference scale, the factor approaches zero, reflecting that insufficient scale cannot generate effective synergy. When no facilities are deployed, the corresponding factor is... Setting this contribution to 1 will not affect the collaborative computing of other facilities. Collaborative performance gain function. The value increases in tandem with the number of facility types and the scale of each facility deployed at the composite site, reflecting the economies of scale effect of composite configuration.

[0114] This step provides a quantitative basis for assessing the positive value of composite sites by quantifying the sharing effects of land conservation and scale synergy, avoiding the one-sidedness of simply assuming that centralized facilities are always better than decentralized ones.

[0115] In one embodiment of the present invention, step S3 includes:

[0116] When multiple types of facilities are deployed simultaneously at candidate location j, interference arises between users of different facilities in terms of space usage, time slot occupancy, and management coordination. This leads to a decline in service quality and an increase in operating costs, demonstrating the negative effect of coordinated facility deployment. The service interference loss function for candidate location j is defined as follows:

[0117] ;

[0118] in For facility type k and The basic interference coefficient between different facility types reflects the degree of service interference. Specifically, the interference coefficient between parking spots and shared stations should be considered in the configuration. The highest level reflects the spatial competition and interference between private vehicle parking and shared vehicle borrowing and returning, as well as the management complexity caused by different user types, and the interference coefficient between parking spots and charging stations. Centered, reflecting the decreased turnover rate caused by charging vehicles occupying parking spaces for extended periods, and the interference coefficient between shared charging stations and charging piles. At the lowest level, this reflects that the charging of shared vehicles is an internal operational activity that can be optimized through scheduling to reduce interference; for example, ;

[0119] The capacity pressure factor, characterizing the nonlinear characteristic of a sharp increase in disturbance intensity when the total scale of facilities at a composite site approaches the site's capacity limit, is defined as:

[0120] ;

[0121] in This represents the maximum space utilization rate of the site, and can be set to 0.70. This is the capacity pressure index, which can take a value of 2.0. The capacity pressure threshold can be set to 0.55. This factor uses a piecewise power function; when the ratio of the facility's occupied area to the available site area is below the threshold of 0.55, the capacity pressure factor remains at a low level. When the ratio exceeds 0.55, the capacity pressure factor increases according to a power function. The closer the ratio is to the upper limit of 0.70, the greater the disturbance intensity. When the ratio reaches 0.70, the factor value is about 1.0, which describes the deterioration of spatial competition when the site is close to saturation.

[0122] The demand coupling factor is calculated as follows:

[0123] ;

[0124] in The demand for facility type k within the service area of ​​candidate location j. The service capacity of facility type k units can be quantified based on the number of people served per day for each facility type. This is the demand pressure coefficient, which can be set to a value of 0.6. To prevent extremely small positive numbers with a denominator of zero;

[0125] The first term of the demand coupling factor is the ratio of the normalized minimum to the maximum value of various types of demand. The closer the relative scales of different types of demand are to 1, the closer this ratio is to 1, indicating a high degree of spatiotemporal overlap and increased interference intensity. When a certain type of demand is significantly lower than others, this ratio approaches 0, indicating a staggered demand distribution and lower interference intensity. The second term is the maximum value of demand saturation. When the demand for a certain type of facility exceeds its service capacity, overload occurs. The more severe the overload, the greater the interference intensity. The terms in parentheses reflect the amplification effect of demand pressure on interference. Service Interference Loss Function The value increases nonlinearly with the increase of space utilization and demand load rate of composite sites, reflecting the congestion cost effect of composite configuration;

[0126] This step provides a quantitative basis for assessing the negative costs of composite sites by quantifying the interference effects of two dimensions: space competition and demand overload, thus correcting the optimistic assumption that service interference is ignored.

[0127] In one embodiment of the present invention, step S4 includes:

[0128] The synergy performance gain function is converted into a synergy performance gain index, and its expression is as follows:

[0129] ;

[0130] in For the synergistic cost-saving rate, a value of 0.25 can be used, indicating that the composite configuration can save 25% of the cost of repeated construction. The basic resource allocation for facility type k. This refers to the amount of resources allocated per unit of scale. and It reflects the resource allocation needs for facility construction, including quantifiable technical parameters such as the amount of site leveling work, the scale of drainage facilities, the number of lighting equipment, and the complexity of the signage system;

[0131] The service interference loss function is converted into an interference effectiveness loss metric, the expression of which is:

[0132] ;

[0133] in The unit service capacity coefficient reflects the service throughput characteristics of a facility and is determined by operational data such as the facility's average daily service users and peak-hour utilization rate. The interference loss rate can be set to 0.15, which quantifies the impact of spatial competition and time-period conflicts between different service types on service throughput under composite configuration.

[0134] The facility configuration efficiency index for candidate location j is defined as the difference between the synergistic efficiency gain index and the interference efficiency loss index, expressed as:

[0135] ;

[0136] Facility configuration efficiency index A positive value indicates that the positive effect of the composite configuration outweighs the negative effect, and the candidate location is suitable for deploying composite sites. This is the facility configuration efficiency index. A negative value indicates that the interference loss exceeds the synergistic gain, and the candidate location should be equipped with a single facility or none at all. By converting the synergistic gain function and the interference loss function into a unified performance evaluation metric, the comparability of synergistic effects and interference effects is achieved. The facility configuration performance index serves as a core decision-making indicator, guiding the trade-off between composite and single configurations.

[0137] In one embodiment of the present invention, step S5 includes:

[0138] Establish a multi-objective optimization model, with the decision variables being: and , respectively, indicate whether a facility of type k and its scale are deployed at candidate location j;

[0139] The optimization model contains two main objective functions. The first objective is to maximize the overall system configuration efficiency, and the objective function is:

[0140] ;

[0141] This objective is to optimize the spatial configuration of composite sites and single sites by balancing the synergistic gains and interference losses of each candidate location.

[0142] The second objective is to maximize the demand-weighted coverage ratio, and the objective function is:

[0143] ;

[0144] Where N is the total number of grid cells. As an auxiliary covering variable, this variable is defined by constraints as follows:

[0145] ;

[0146] in For indicator functions, The maximum service radius for facility type k, specifically, is [value to be filled in]. The service radius of parking spots (600m) and collaborative stations (800m) reflects the greater mobility of collaborative vehicles, while the service radius of charging piles (500m) reflects users' higher requirements for charging convenience. When grid i is within the service radius of at least one deployed facility... A value of 1 indicates that the content has been overwritten; otherwise... A value of 0 indicates no coverage, and this objective ensures that high-demand areas receive priority service.

[0147] The third objective is to minimize the total cost of facility construction, and the objective function is:

[0148] ;

[0149] in This represents the basic resource allocation for facility type k. Let k be the unit size resource allocation for facility type. The site development cost for candidate location j is determined based on land prices and terrain conditions;

[0150] The constraints of the optimization model include the following seven items:

[0151] Constraint 1 is a lower bound constraint on the size:

[0152] ;

[0153] in Let k be the minimum economic scale for facility type k, and its specific value is... The minimum scale for parking spots is 20 parking spaces, the minimum scale for shared stations is 10 vehicles, and the minimum scale for charging piles is 5 piles. This constraint ensures that the deployed facilities meet the minimum operational threshold.

[0154] Constraint 2 is an upper bound constraint on size:

[0155] ;

[0156] in The maximum allowable size for facility type k, specifically, is [value to be filled in]. The maximum number of parking spaces is 200, the maximum number of shared stations is 100 vehicles, and the maximum number of charging piles is 50. This constraint prevents the scale of single-point facilities from becoming too large.

[0157] Constraint 3 is a spatial capacity constraint:

[0158] ;

[0159] Ensure that the total area occupied by facilities at the composite site does not exceed the maximum utilization rate of the available site area, of which The value is 0.70;

[0160] Constraint 4 is a core requirement coverage constraint, implemented using a soft constraint approach by adding a penalty term to the objective function.

[0161] ;

[0162] in The core demand area set is defined as the demand potential index. The top 30% of grid cells, With a penalty coefficient of 100, this constraint encourages high-demand areas to be served by every type of facility through a penalty term, but allows some core grids to be not fully covered to avoid the model being unsolvable, which meets the flexible requirements of actual planning.

[0163] Constraint 5 is the budget constraint:

[0164] ;

[0165] in The basic resource allocation for facility type k. Let k be the unit size resource allocation for facility type. B represents the site development cost for candidate location j, determined based on land prices and terrain conditions; B is the total budget ceiling, which limits the total investment in facility construction.

[0166] Constraint 6 is the minimum spacing constraint for similar facilities, for any candidate location pair ,if ,but ;in For candidate positions j and The Euclidean distance between them The minimum allowable spacing for facility type k is given by the value specified in the original text. The minimum spacing between parking spots is 400m, the minimum spacing between shared stations is 500m, and the minimum spacing between charging piles is 350m. This constraint only applies to similar facilities to prevent excessive concentration of similar facilities from causing overlapping service areas. Different types of facilities can be very close to each other or even located in the same place.

[0167] Constraint 7 is a coverage indication constraint:

[0168] ;

[0169] This constraint defines the covering variable. With the layout of decision variables The logical relationship.

[0170] like Figure 3 As shown, in one embodiment of the present invention, step S6 includes:

[0171] The multi-objective genetic algorithm NSGA-II was used to solve the above multi-objective integer programming model;

[0172] During the algorithm initialization phase, an initial population of size 100 is generated, with each individual representing a complete facility layout scheme. Chromosomes are encoded using integers, with a length equal to the number of candidate positions (n) multiplied by the number of facility types (3). Each gene position represents the scale value of the corresponding candidate position and facility type. Indicates no deployment Positive integers represent the corresponding deployment scale. The initial population, with each gene locus ranging from 0 to the maximum size corresponding to the facility type. Randomly generated between;

[0173] The fitness of each individual in the initial population is evaluated, and the specific calculation process is as follows:

[0174] First, the decision variables are analyzed from the chromosome coding of the individual. For candidate position j and facility type k, if the value of the corresponding gene locus in the chromosome... Then let And record the scale ,like Then let ;

[0175] Then, the evaluation function value for each candidate position is calculated sequentially. For each candidate position j, its decision variables are... and Substitute the synergistic efficiency gain function formula defined in step S2 to calculate the dispersed occupied area. Maximum Shared Area Composite area Land intensive use factors and scale synergy factor Finally, the cooperative efficiency gain value of the candidate position is obtained. ;

[0176] Next, substitute the decision variables into the service disturbance loss function formula defined in step S3 to calculate the capacity stress factor. and demand coupling factor Finally, the service interference loss value of the candidate location is obtained. ;

[0177] Then and Substitute the facility configuration efficiency index formula defined in step S4 to calculate the synergistic efficiency gain index. and interference effectiveness loss index The configuration efficiency value of the candidate position is obtained. ;

[0178] After traversing all candidate positions and completing the above calculations, the configuration performance value of each candidate position is... Substitute the first objective function formula in step S5 to calculate the overall system configuration efficiency. If the objective function uses a penalty term, then the penalty term is calculated and corrected accordingly. ;

[0179] Based on decision variables Determine the coverage of each grid cell i. If the distance between grid i and any deployed facility is within the service radius of that facility, then let... otherwise The demand potential index of each grid and covered variables Substitute into the second objective function formula to calculate the demand-weighted coverage rate. ;

[0180] decision variables and Substitute into the third objective function formula to calculate the total cost of facility construction. ;

[0181] After completing the above calculations, the three objective function values ​​for this individual are... As its coordinates in the target space;

[0182] The algorithm enters an iterative process, with each generation including four operations: selection, crossover, mutation, and elite retention. The selection operation uses a tournament selection method with a tournament size of 3. Three individuals are randomly selected from the current population for comparison, and the individual with the lowest non-dominant level is chosen. If the non-dominant levels are the same, the individual with the largest crowding distance is chosen as the parent. The crossover operation uses a two-point crossover with a crossover probability of 0.85. Two parent individuals are randomly selected, and two crossover points are randomly chosen on the chromosomes. The gene segments between the two crossover points are exchanged to generate two offspring individuals. The mutation operation uses uniform mutation with a mutation probability of 1 / (3n). Each gene locus is mutated with a mutation probability, and the upper and lower bounds of the facility type size corresponding to that gene locus are used during mutation. Randomly generate new values ​​within the range;

[0183] Constraint checks and repairs are performed on the generated offspring individuals. For size constraints, if a gene locus has a non-zero value but is less than the minimum economic size... Then it will be corrected to Regarding space capacity constraints, if the actual area occupied by an individual composite site... Exceeding the available capacity of the site Then, the scale of each facility is reduced proportionally until the constraints are met. For spacing constraints, if the distance between two candidate locations is... If both have type k facilities deployed, then the size of one of the type k facilities will be randomly set to 0.

[0184] The fitness assessment process was repeated for the repaired offspring, and the decision variables were analyzed from the chromosomes to calculate the synergistic efficacy gain for each candidate position. Service Interference Loss and configuration performance Then calculate the values ​​of the three objective functions. , and The parent and offspring populations are merged into a temporary population of size 200. A fast non-dominated sort is performed on this temporary population. Individuals are then divided into Pareto fronts of different levels based on their dominance relationships. The first front represents the non-dominated solution, the second front represents the non-dominated solution after removing the first front, and so on. For each front, the crowding distance is calculated. Crowding distance represents the average distance between an individual and its neighbors in the target space. Specifically, for each target dimension, individuals within the front are sorted according to the target value. The crowding distance for boundary individuals is set to infinity, and the crowding distance for intermediate individuals is the normalized difference between the target values ​​of their immediate and immediate neighbors in that target dimension. The total crowding distance for each individual is obtained by summing the crowding distances across all target dimensions. The temporary population is then sorted according to non-dominated level and crowding distance, prioritizing individuals with lower non-dominated levels and those with larger crowding distances within the same level. The top 100 individuals are selected to enter the next generation population.

[0185] The maximum number of iterations for the algorithm is set to 300. The convergence criterion is that the rate of change of the hypervolume index of the non-dominated solution set is less than 0.5% over 50 consecutive iterations. The hypervolume index is calculated using a reference point... Based on this, calculate the target spatial volume covered by the Pareto front. This is 1.5 times the total budget ceiling. The rate of change of the super-volume index over 50 consecutive generations... When the algorithm is considered to have converged, the iteration is terminated prematurely.

[0186] The algorithm outputs a Pareto optimal solution set, which consists of all individuals in the first Pareto front at the time of algorithm termination. Each solution includes the combination of facility types and scale configurations for each candidate location, specifically represented as a decision variable matrix. Simultaneously record the three objective function values ​​of the solution. and the configuration performance value of each candidate position This provides decision-makers with a variety of options.

[0187] In one embodiment of the present invention, step S7 includes:

[0188] The facility configuration efficiency index values ​​of each candidate location are statistically analyzed from the Pareto optimal solution set. Candidate locations with positive facility configuration efficiency index values ​​are selected, and their comprehensive demand potential index, available site area, point of interest function mixing degree and public transport accessibility characteristics are extracted.

[0189] The definition of point-of-interest (POI) functionality blending is three types of requirements. coefficient of variation ,in Standard deviation The coefficient of variation is the mean; a smaller coefficient indicates a more balanced demand and a higher degree of functional mixing. Public transport accessibility characteristics are measured using bus stop density. The ratio to the regional average, and the coverage of public transport routes. As an indicator;

[0190] Based on the extracted features, cluster analysis was used to determine the characteristic thresholds for suitable areas for composite configuration, identifying high-demand-density areas, areas surrounding public transport hubs, and mixed-use zones. The k-means clustering algorithm was used to divide candidate locations with positive facility configuration efficiency index values ​​into two categories: strongly suitable areas and moderately suitable areas. Statistical analysis was used to identify the characteristic thresholds for strongly suitable areas: the comprehensive demand potential index. And usable area The area is characterized by high demand density (square meters), and the density of bus stops is also high. Greater than 1.5 times the regional average and with bus route coverage The route is located near a public transport hub; the coefficients of variation for the three types of demand. It is a mixed-function area;

[0191] Based on the facility configuration efficiency index value and the identified feature thresholds, the facility type configuration and deployment scale for each candidate location are output. Candidate locations with a positive facility configuration efficiency index value and meeting the feature thresholds are recommended to be configured as composite sites containing parking spots and charging piles, or all three types of facilities. Candidate locations with a negative facility configuration efficiency index value or close to zero are recommended to be configured with a single type of facility. The specific configuration rule is: configuration efficiency value... And the demand potential index The candidate locations are configured with composite sites for three types of facilities, and the efficiency values ​​are configured accordingly. But demand potential index Candidate locations with a performance value between 0.4 and 0.7 are selected for the deployment of dual-mode composite sites for parking spots and charging piles, with configured efficiency values. The candidate locations are configured with single facilities based on demand, and the facility type with the highest demand percentage is selected.

[0192] This step establishes quantitative configuration decision rules through facility configuration efficiency index and cluster analysis, solves the technical problem of which locations are suitable for composite configuration, and provides quantitative basis for planning decision-makers to make differentiated configurations.

[0193] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A collaborative optimization method for the layout of electric bicycle facilities based on multi-dimensional features, characterized in that, include: S1. Based on road network density characteristics, point of interest distribution characteristics, public transport station accessibility characteristics and population density characteristics, obtain the comprehensive demand potential index of candidate locations and decompose it into parking facility demand, shared facility demand and charging facility demand. S2. Based on the degree of resource reuse, land intensive use and scale synergy among different facility types, construct a synergy efficiency gain function for each candidate location; The synergistic performance gain function is: ； Where j is the candidate position index; k and K is the facility type index; K is the set of facility types, including parking spots, shared stations, and charging stations. The decision variable is defined as whether facility type k is deployed at candidate location j. For facility type k and The basic synergy coefficient between them reflects the degree of resource reuse; As a factor for intensive land use; The scale synergy factor is calculated as follows: ； In the formula, The scale of facility type k deployed for candidate location j; This is a reference size for facility type k; Total number of facility types; S3. Based on the degree of service interference, capacity pressure level and coupling degree of parking facility demand, shared facility demand and charging facility demand among different facility types, construct a service interference loss function for each candidate location. The service interference loss function is: ； Where j is the candidate position index; k and K is the facility type index; K is the set of facility types; To define decision variables; For facility type k and The basic interference coefficient between them reflects the degree of service interference; Capacity pressure factor; The demand coupling factor is calculated as follows: ； In the formula, The demand for facility type k within the service area of ​​candidate location j. For the reference size of facility type k, The scale of facility type k deployed for candidate location j; The service capacity per unit size for facility type k; This is the demand pressure coefficient; To prevent extremely small positive numbers with a denominator of zero; S4. Construct the facility configuration efficiency index for each candidate location based on the collaborative efficiency gain function and the service interference loss function; S5. With the objectives of maximizing the overall system configuration efficiency, maximizing the demand-weighted coverage rate, and minimizing the total construction cost, and with constraints of scale, spatial capacity, coverage, budget, and spacing, a multi-objective programming model is established; wherein the overall system configuration efficiency is the sum of the facility configuration efficiency indices of each candidate location, and the weight of the demand-weighted coverage rate is the comprehensive demand potential index. S6. Solve the multi-objective programming model using a multi-objective genetic algorithm to obtain the Pareto optimal solution set; S7. Identify suitable regional characteristics of composite and single configurations from the Pareto optimal solution set based on the facility configuration efficiency index value, and output the facility type configuration and deployment scale of each candidate location.

2. The method for collaborative optimization of electric bicycle facility layout based on multi-dimensional features according to claim 1, characterized in that, Step S1 obtains a comprehensive demand potential index for candidate locations based on road network density characteristics, point-of-interest distribution characteristics, public transport stop accessibility characteristics, and population density characteristics, including: The study area was divided into square grid cells. For each grid cell, the road network density features, point of interest distribution features, public transport station accessibility features, and population density features were extracted. After normalizing each feature, a weighted summation method was used to calculate the comprehensive demand potential index. The road network density features include road density, non-motorized vehicle lane coverage, and road network connectivity index; the point of interest distribution features include residential area density, commercial area density, office area density, and educational facility density; the public transport stop accessibility features include public transport stop density, public transport route coverage, and average distance between public transport stops; and the population density features include resident population density and employed population density.

3. The method for collaborative optimization of electric bicycle facility layout based on multi-dimensional features according to claim 2, characterized in that, Step S1 decomposes the comprehensive demand potential index into parking facility demand, shared facility demand, and charging facility demand, including: For each candidate location, the comprehensive demand potential index of each grid cell within the service range is weighted and summed using the distance decay function to obtain the total demand within the service range of that candidate location. Based on the density of residential areas, commercial areas, office areas, and educational facilities in the grid unit where the candidate location is located, a demand decomposition rule is established and the total demand is decomposed into three types of facility demand. Among them, the density of residential areas dominates the demand for parking facilities, the density of commercial areas dominates the demand for shared facilities, and the density of office areas and educational facilities jointly affect the demand for charging facilities.

4. The method for collaborative optimization of electric bicycle facility layout based on multi-dimensional features according to claim 1, characterized in that, Step S4 includes: The synergy performance gain function is converted into a synergy performance gain index, and its expression is as follows: ； Where j is the candidate location index; k is the facility type index; and K is the set of facility types, including parking spots, shared stations, and charging piles. The synergistic performance gain function; To define decision variables; The scale of facility type k deployed for candidate location j; The basic resource allocation for facility type k; This is a reference size for facility type k; This refers to the amount of resources allocated per unit of scale. To improve the collaborative savings rate; The service interference loss function is converted into an interference effectiveness loss metric, the expression of which is: ； in, For service interference loss function; The service capacity per unit size for facility type k; Service capacity factor per unit; To compensate for the loss rate; The facility configuration efficiency index is defined as follows: ； in, As a synergistic performance gain indicator, This is an indicator of interference effectiveness loss.

5. The method for collaborative optimization of electric bicycle facility layout based on multi-dimensional features according to claim 2, characterized in that, In step S5, the objective function of the multi-objective programming model is: ； ； in, L represents the overall configuration efficiency; L is the set of candidate positions; j is the index of the candidate position. The facility configuration efficiency index for candidate location j; The coverage is a demand-weighted ratio; N is the total number of grid cells; i is the grid cell index. The comprehensive demand potential index for grid cell i; This is an auxiliary coverage variable, indicating whether grid cell i is covered.

6. The method for collaborative optimization of electric bicycle facility layout based on multi-dimensional features according to claim 5, characterized in that, The constraints of the multi-objective programming model include: Size constraints Limit the deployment scale of each facility type; Space capacity constraints Limit the space occupancy rate of composite sites; Coverage constraints Define the sufficiency of grid coverage; Budget constraints The total resource investment in the construction of control facilities shall not exceed the budget limit B; Spacing constraints Avoid excessive concentration of similar facilities; Where K is the set of facility types, and k is the facility type index. and Let these be the minimum and maximum sizes of facility type k, respectively. Let j be the actual area occupied by the composite site at candidate location j. For usable area, To maximize the use of space, For candidate position index, Let j be the distance from the candidate position j to the center of grid cell i. Let B be the maximum service radius of facility type k, and B be the total budget ceiling. The basic resource allocation for facility type k. To define decision variables; Let k be the unit size resource allocation for facility type. The scale of facility type k deployed for candidate location j. The site development cost for candidate location j. For indicator functions, For candidate positions j and The distance between them The minimum allowable spacing for facility type k.

7. The method for collaborative optimization of electric bicycle facility layout based on multi-dimensional features according to claim 1, characterized in that, In step S6, the multi-objective genetic algorithm uses the NSGA-II algorithm; chromosomes are encoded with integers to represent the scale configuration of each facility type at each candidate location, with a gene value of 0 indicating no facility of that type, and a positive integer indicating the deployment of facilities of the corresponding scale; the optimization strategy combines tournament selection, two-point crossover, and uniform mutation, and fitness is evaluated using fast non-dominated sorting and crowding distance indicators; the algorithm prioritizes individuals with low non-dominated levels and large crowding distances; the optimization stopping condition is that the rate of change of the hypervolume index of the non-dominated solution set for M consecutive generations is less than 1%. .

8. The method for collaborative optimization of electric bicycle facility layout based on multi-dimensional features according to claim 2, characterized in that, Step S7 includes: The facility configuration efficiency index values ​​for each candidate location are statistically analyzed from the Pareto optimal solution set. Candidate locations with positive facility configuration efficiency index values ​​are selected, and their comprehensive demand potential index, available site area, point of interest functional mixing degree, and public transport accessibility characteristics are extracted. Based on the extracted features, cluster analysis is used to determine the feature thresholds of suitable areas for composite configuration, and to identify high-demand density areas, areas around public transport hubs, and mixed-use areas. Based on the facility configuration efficiency index value and the identified feature threshold, the facility type configuration and deployment scale of each candidate location are output. Candidate locations with a positive facility configuration efficiency index value and meeting the feature threshold are deployed with composite sites, while other candidate locations are deployed with single-type facilities.

Images