A method for modeling activity chain time allocation behavior considering travel cost

Abstract

This invention discloses a method for modeling activity chain time allocation behavior that considers travel costs. The method includes acquiring vehicle travel trajectories, calculating the inherent utility of activities based on the influence of activity-related attributes and individual attributes on individual preferences, calculating the travel-related negative utility of activities based on the influence of necessary travel-related attributes on individual preferences, and constructing a utility function that comprehensively considers both the inherent utility and travel-related negative utility. Based on utility maximization theory and time conservation theory, a joint decision-making model for activity selection and time allocation is established. Statistical methods are used to solve the constructed model to obtain the optimal activity time allocated to each destination, thereby maximizing the overall activity utility. This method can more accurately characterize the time allocation process of individuals between different activity locations, improving the model's applicability and prediction accuracy to real-world travel behavior.

Description

Technical Field

[0001] This invention relates to the field of traffic behavior modeling, and more specifically to a method for modeling the time allocation behavior of activity chains that takes into account travel costs. Background Technology

[0002] The allocation of activity time has attracted widespread attention in fields such as psychology, economics, and transportation science. Existing time allocation models are generally based on the assumption that individuals (and their families) will allocate time rationally across all activities to maximize total utility. However, because travel and activity behaviors are intertwined and mutually influential, current research in time allocation modeling still faces two key challenges:

[0003] The existing models fail to consider travel time. They typically focus solely on the time consumed by the activity itself during time allocation, neglecting the travel time required to reach the activity location. However, in daily life, individuals spend not only on the activity itself but also a significant portion of their time commuting and traveling. Ignoring travel time will lead to biased estimates of actual activity time, thus reducing the model's accuracy.

[0004] Existing models often overlook the complex interaction between travel and activity behavior. A close interaction exists between travel behavior and activity selection. In the destination selection process, accessibility significantly impacts individual decisions. If the route to an activity location is inconvenient, individuals may abandon certain activities due to excessive travel costs. However, existing models often neglect the travel time connecting various activities and the accessibility of different activity locations, thus failing to accurately depict the dynamic relationship between travel and activities.

[0005] To address the shortcomings of traditional methods, this invention proposes a behavioral modeling approach for activity chain time allocation that considers travel costs. This model introduces travel cost as a key variable. Travel cost is primarily determined by the travel time connecting various activities; as travel cost increases, the utility an individual gains from the activity decreases accordingly, thus more realistically reflecting the trade-offs individuals make when choosing activities. Furthermore, we incorporate travel time into the model, considering not only the time consumed by the activity itself but also the time factor of reaching the destination in the objective function and constraints. This improvement enables the model to more accurately describe an individual's time allocation across various activities, enhancing its ability to characterize real-world behavior. Summary of the Invention

[0006] The purpose of this invention is to provide a method for modeling the time allocation behavior of activity chains that considers travel costs. This method differs from traditional activity time allocation models. This invention not only introduces travel costs as a key variable into the objective function and constraints, but also further incorporates travel costs. The model can more realistically reflect the time consumption faced by individuals in the process of activity selection, thereby improving its ability to describe actual travel behavior.

[0007] To achieve the above functions, this invention designs a behavior modeling method for activity chain time allocation that considers travel costs. For the travel behavior of an individual vehicle, the following steps S1-S4 are executed to complete the activity time allocation of the vehicle at different destinations:

[0008] Step S1: Use the AVI detector to obtain the vehicle's travel trajectory, including its driving path and multiple possible destinations, and construct a travel trajectory dataset containing all potential activity locations;

[0009] Step S2: For each activity and vehicle, calculate the inherent utility of each activity based on the influence of activity-related attributes and individual attributes on individual preferences. Calculate the travel-related utility of each activity based on the influence of travel-related attributes necessary to reach the activity location on individual preferences. Construct a utility function that comprehensively considers the inherent utility and travel-related utility of each activity to describe the individual's preferences in different activity choices.

[0010] Step S3: Based on the utility maximization theory and the time conservation theory, establish a joint decision-making model for activity selection and time allocation. The joint decision-making model is used to characterize an individual's activity selection among discrete options and quantify the individual's time allocation for each selected activity.

[0011] Step S4: Use statistical methods to solve the constructed joint decision-making model to obtain the optimal activity time allocated to each destination, so as to maximize the overall activity utility and complete the allocation of vehicle activity time to different destinations.

[0012] Beneficial effects: Compared with the prior art, the advantages of the present invention include:

[0013] 1. Introducing the travel time factor. Traditional time allocation models only focus on the allocation of activity time, without considering the travel time to the activity location. This invention introduces the travel time factor and incorporates it into the objective function and constraints, enabling the model to simultaneously consider the time allocation of travel and activities when estimating individual activity time allocation, thereby improving the accuracy of time allocation.

[0014] 2. The invention incorporates travel cost factors. Travel costs are primarily determined by the travel time connecting various activities; as travel costs increase, the utility an individual derives from the activity decreases accordingly. Furthermore, this invention considers the complex relationship between travel negative utility and activity time and type. Compared to traditional methods that only consider the utility of the activity itself, this improvement is more consistent with the actual decision-making process. Attached Figure Description

[0015] Figure 1 This is a flowchart of a method for modeling the behavior of activity chain time allocation considering travel costs, provided by an embodiment of the present invention;

[0016] Figure 2 This is a schematic diagram of the activity time allocation between adjacent AVI detectors according to an embodiment of the present invention. Detailed Implementation

[0017] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0018] This invention provides a method for modeling the activity chain time allocation behavior that considers travel costs. For the travel behavior of an individual vehicle, the following steps S1-S4 are executed to complete the activity time allocation of the vehicle at different destinations:

[0019] Step S1: Use the AVI detector to obtain the vehicle's travel trajectory, including its driving path and multiple possible destinations, and construct a travel trajectory dataset containing all potential activity locations;

[0020] In step S1, the AVI detector is used to obtain the travel trajectory of the vehicle, and the nth vehicle is recorded as being detected by the detector. The recorded timestamp is The nth vehicle is detected. The recorded timestamp is According to the law of conservation of time, the sum of the vehicle's travel time between the two AVI detectors and its destination dwell time (activity duration) must be less than the time budget, which is the sum of the vehicle's travel time between the detectors. and The difference in the recorded timestamps.

[0021] Step S2: For each activity and vehicle, calculate the inherent utility of each activity based on the influence of activity-related attributes and individual attributes on individual preferences. Calculate the travel-related utility of each activity based on the influence of travel-related attributes necessary to reach the activity location on individual preferences. Construct a utility function that comprehensively considers the inherent utility and travel-related utility of each activity to describe the individual's preferences in different activity choices.

[0022] The utility function described in step S2, which comprehensively considers the inherent utility and travel negative utility of each activity, is as follows:

[0023]

[0024] In the formula, U(T) represents the utility function, and K represents the number of potential activity locations between two adjacent AVI detectors, that is, the total number of all activities that an individual may participate in; This indicates the influence of activity-related attributes and individual attributes on individual preferences. These attributes include factors such as activity type, arrival time, and individual socioeconomic attributes. This indicates the impact of travel-related attributes necessary to reach the activity location on individual preferences; these attributes include factors such as travel time and travel cost. Indicates the duration of participation in activity k; γ k α k ρ k and η k It is an adjustable parameter used to characterize the specific influence of various practical factors on utility; among them, γ k Assign parameters to time; α k α is a saturation parameter, representing the degree of diminishing marginal utility of activity k, where 0 < α. k <1;ρ k The parameter representing the influence of activity type, 0 < ρ k <1;η k The parameter representing the influence of the activity duration, -1 < η k <0.

[0025] The utility function U(T) is the sum of the utility derived from allocating time at various activity locations between two adjacent AVI detectors. This utility function can describe not only the case where there are multiple potential activity locations between two adjacent AVI detectors, but also the special case where there are no activity locations (i.e., k=0). When k=0, the individual's total utility simplifies to... At this point, the utility is determined solely by the negative utility resulting from the journey between the two detectors, which aligns with actual travel behavior.

[0026] Taking into account the inherent utility of each activity and the negative utility of travel, the utility function can be decomposed into the following two product terms:

[0027]

[0028] and

[0029]

[0030] in, This represents the inherent utility of activity k, and the positive utility an individual gains from participating in activity k. This represents the baseline utility of activity k. To ensure that this baseline utility is always positive, it is modeled in exponential form, α. k Represents the saturation parameter; γ k ρ represents the time allocation parameter. k η represents the influence parameter of the activity type. k The parameter representing the impact of the activity duration, Indicates the duration of participation in activity k, ∈ k This represents the directly observed random factors that affect the baseline utility of activity k. The influence of activity-related attributes and individual attributes on individual preferences is represented by a function composed of observable features related to the activity, specifically determined by the individual's socioeconomic attributes and the characteristics of the activity itself. Its mathematical expression is as follows:

[0031]

[0032] In the above formula, activity utility is influenced by both individual socioeconomic attributes and the characteristics of the activity itself. Specifically, the attributes of the activity location determine the suitability of different types of activities, thus affecting individual preferences. Simultaneously, individual characteristics (such as age and income level) play a significant role in activity demand, with significant differences among different groups in their choice of activity type and location. Furthermore, the time factor cannot be ignored; arrival time affects the activity experience and convenience. Peak hours may reduce utility due to congestion, while off-peak hours offer a relatively optimized activity environment. Similarly, the date characteristics of the activity (such as weekdays or weekends) moderate individual behavioral patterns, determining whether they prefer work, commuting, or leisure activities.

[0033] In the above formula, x location x age x income x arrival time x weekday The weights;

[0034] x location These variables represent the characteristics of the activity location; different locations can accommodate different types of activities, and therefore their corresponding activity utility also varies. For example, commercial areas, entertainment centers, and office areas may have different weights in an individual's utility assessment;

[0035] x ageThis represents the individual's age variable; individuals of different age groups have significant differences in their preferences for activity locations. For example, children may prefer places like amusement parks or schools, so these places have higher utility for children, while adults are more likely to go to their workplaces, so their utility value at the workplace is higher.

[0036] x income This indicates an individual's income level; high-income groups are generally more inclined to participate in high-consumption activities, such as high-end dining, leisure and entertainment, or vacations, and therefore these activities are more attractive to this group, with correspondingly higher utility values;

[0037] x arrival time The arrival time represents the time it takes for an individual to arrive at the activity location; the experience and convenience of an activity are often affected by the arrival time. For example, peak hours may bring higher congestion costs or waiting times, thereby reducing the utility of the activity, while off-peak hours may provide a more comfortable activity environment and improve the overall utility.

[0038] x weekday This indicates the date on which the activity takes place; that is, whether it is a workday. Workday activities are usually based on work and commuting, so individuals are more inclined to go to their workplaces, while on weekends they are more inclined to participate in leisure and entertainment activities, such as shopping malls, amusement parks or cultural activity centers.

[0039] at the same time It also reflects the non-linear relationship between activity utility and activity time. Specifically, as activity time increases, activity utility does not increase linearly, but follows the law of diminishing marginal utility. That is, the longer an individual invests in an activity, the slower the increase in their additional utility. k This reflects the degree of saturation (i.e., the degree of diminishing marginal utility);

[0040] γ k The time allocation pattern influences whether an individual adopts a corner solution (i.e., participating in only one activity) or an internal solution (i.e., allocating time among multiple activities). Specifically, when γ... k When γ ≠ 0, the individual will not allocate any time to activity k. And when γ k When the value is 0, it means that the individual will at least invest some time in the activity.

[0041] Furthermore, this form of utility can adapt to various time allocation scenarios, depending on the specific circumstances. and α k The value of k; if a specific activity k has a high benchmark utility (relative to all other activities), and its saturation parameter α kA value close to 1 indicates that the individual has an extremely high preference for the activity and minimal diminishing marginal utility, meaning their "saturation" with the activity is extremely low; therefore, the individual will allocate almost all their time to this activity. On the other hand, when the individual's baseline utility across all activities... They are approximately equal, and α k When the value is small, individuals tend to distribute their time evenly among multiple activities rather than focusing on one activity;

[0042] By adjusting α k With γ k The value of can not only characterize an individual's single activity preference, but also reflect diverse time allocation behaviors, and has strong flexibility and applicability.

[0043] Product terms in utility function U(T) This represents the travel negative utility of activity k, used to characterize the impact of travel costs on activity utility. An individual's decision to participate in an activity is influenced not only by the inherent utility of the activity itself, but also by the travel costs required to reach the activity location. For example, when traveling to an activity location requires a long travel time or there is severe traffic congestion, an individual may choose to abandon the activity and participate in other activities instead. Therefore, introducing a travel burden decay coefficient helps to more realistically reflect an individual's preference for activities. As travel time or distance increases, this decay coefficient decreases, thereby reducing the overall utility of the activity.

[0044] To ensure that the basic utility is positive, an exponential function is used, where α k Represents the saturation parameter; γ k ρ represents the time allocation parameter. k η represents the influence parameter of the activity type. k The parameter representing the impact of the activity duration, This indicates the duration of participation in activity k. The mathematical expression for the influence of travel-related attributes necessary to reach location k on individual preferences is as follows:

[0045]

[0046] In the formula, x distance Indicates travel distance, x numofsignallights Indicates the number of traffic lights along the route. x distance x numofsignallights The weight.

[0047] However, the negative utility of travel depends not only on the direct cost of travel, but also on the duration and type of the activity; to effectively characterize these factors, the model introduces the parameter ρ. k and ηk It is used to describe changes in travel burden based on activity characteristics, enabling the model to more accurately depict the relationship between activities and travel costs, thereby more realistically describing the individual's decision-making process;

[0048] The negative utility of travel varies with the duration of the activity; for example, the negative utility of a 1-hour commute is relatively small for long-duration activities (such as an 8-hour workday), while the same 1-hour commute incurs a more significant burden for short-duration activities (such as a 1-hour workday). To more accurately describe the negative correlation between activity duration and travel utility, a negative parameter η is introduced. k (i.e., -1 < η) k <0), thus better depicting the impact of changes in travel time cost with activity duration;

[0049] The negative utility of travel is also affected by the type of activity; for example, for two activities of the same duration (such as two hours of work and two hours of shopping), if the travel distance for both is the same (i.e., If the values ​​are the same, one might assume that the travel negative utility of these two activities is the same; however, this assumption is not valid.

[0050] Work, as a rigid activity, has low flexibility, meaning individuals find it difficult to avoid such activities; while shopping is a flexible activity, giving individuals greater freedom in their travel choices; therefore, for rigid activities like work, let ρ be... k Take the smaller value (i.e., 0 < ρ) k <1) indicates that the negative utility of commuting to work is lower than that of flexible activities such as shopping; through this adjustment, the model can more reasonably reflect real-world behavior, that is, compared to entertainment or leisure activities, the perceived burden of commuting for rigid activities such as work is lower.

[0051] Step S3: Based on the utility maximization theory and the time conservation theory, establish a joint decision-making model for activity selection and time allocation. The joint decision-making model is used to characterize an individual's activity selection among discrete options and quantify the individual's time allocation for each selected activity.

[0052] The joint decision-making model for activity selection and time allocation established in step S3 is as follows:

[0053]

[0054] The joint decision-making model describes the decision-making process of individuals maximizing stochastic utility under time conservation constraints, where the time budget constraint is:

[0055]

[0056] in, This represents the total amount of time an individual can control. This indicates the duration of participation in activity k. Let K represent the travel time, and K represent the number of potential activity locations between two adjacent AVI detectors. Using AVI detectors, the travel trajectory of an individual vehicle can be obtained, including the driving path and multiple possible destinations; therefore, for each individual vehicle, the travel time... This can be considered as known information. Based on this, the above model can be constructed as a solvable convex optimization problem, thereby ensuring the existence and feasibility of the optimal solution.

[0057] Step S4: Use statistical methods to solve the constructed joint decision-making model to obtain the optimal activity time allocated to each destination, so as to maximize the overall activity utility and complete the allocation of vehicle activity time to different destinations.

[0058] In step S4, the joint decision model is solved, and the solution results are as follows:

[0059]

[0060] In the above solution, K represents the number of potential activity locations between two adjacent AVI detectors, M represents the number of activities that the sample vehicle chooses to participate in among the K activities, and 1≤M≤K. Let k represent the duration of participation in activity k, where k = 1, 2, ..., M, ..., K;

[0061] Where, θ k The intermediate variable is represented by the following expression:

[0062]

[0063] V k The deterministic utility of the k-th activity is determined by both the activity-related attributes and the travel attributes required to reach the activity location. The specific expression is as follows:

[0064]

[0065] Where, α k Represents the saturation parameter; γ k ρ represents the time allocation parameter. k η represents the influence parameter of the activity type. k The parameter representing the impact of the activity duration.

[0066] Refer to the diagram showing the activity time allocation between adjacent AVI detectors. Figure 2 By applying the activity chain time allocation behavior modeling method between adjacent detectors proposed in this invention, the activity time of each sample vehicle captured by the detector can be reasonably allocated.

[0067] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A method for modeling the time allocation behavior of activity chains considering travel costs, characterized in that, For individual vehicle travel behavior, perform the following steps S1-S4 to allocate the vehicle's activity time to different destinations: Step S1: Use the AVI detector to obtain the vehicle's travel trajectory, including its driving path and multiple possible destinations, and construct a travel trajectory dataset containing all potential activity locations; Step S2: For each activity and vehicle, calculate the inherent utility of each activity based on the influence of activity-related attributes and individual attributes on individual preferences. Calculate the travel-related utility of each activity based on the influence of travel-related attributes necessary to reach the activity location on individual preferences. Construct a utility function that comprehensively considers the inherent utility and travel-related utility of each activity to describe the individual's preferences in different activity choices. The utility function described in step S2, which comprehensively considers the inherent utility and travel negative utility of each activity, is as follows: ； In the formula, Represents the utility function. This indicates the number of potential activity locations between two adjacent AVI detectors; This indicates the influence of activity-related attributes and individual attributes on individual preferences; This indicates the impact of travel-related attributes necessary to reach the activity location on individual preferences; Indicate participation in the activity Duration; , , and It is a parameter with adjustable values, where, Assign parameters to time, when At that time, the individual will not be active. Allocate any time, when At that time, individuals will be in the activity Allocate a certain amount of time above; The saturation parameter represents the activity. The degree of diminishing marginal utility, 0 < <1; The parameter representing the influence of the activity type, 0 < <1; The parameter representing the impact of the activity duration, -1 < <0; The utility function in step S2, which comprehensively considers the inherent utility and travel disutility of each activity, is decomposed into the following two product terms: ； and ； in, Indicates activity Its solid effectiveness Indicates activity The baseline utility This represents the saturation parameter; Indicates time allocation parameters, The parameter representing the impact of the activity type. The parameter representing the impact of the activity duration, Indicate participation in the activity Duration Indicates the impact of the activity Directly observed random factors of baseline utility Indicates activity The mathematical expression for the influence of relevant attributes and individual attributes on individual preferences is as follows: ； In the formula, Characteristic variables representing the location of the activity; The age variable represents an individual; Indicates an individual's income level; The time when an individual arrives at the activity location; Indicates the date characteristic of the event. , , , , They are respectively , , , , The weights; Step S3: Based on the utility maximization theory and the time conservation theory, establish a joint decision-making model for activity selection and time allocation. The joint decision-making model is used to characterize an individual's activity selection among discrete options and quantify the individual's time allocation for each selected activity. Step S4: Use statistical methods to solve the constructed joint decision-making model to obtain the optimal activity time allocated to each destination, so as to maximize the overall activity utility and complete the allocation of vehicle activity time to different destinations.

2. The activity chain time allocation behavior modeling method considering travel costs according to claim 1, characterized in that, In step S1, the vehicle's travel trajectory is obtained using an AVI detector, and the first... One vehicle was detected The recorded timestamp is , No. One vehicle was detected The recorded timestamp is According to the law of conservation of time, the sum of the vehicle's travel time between the two AVI detectors and its destination dwell time must be less than the time budget, which is the sum of the time the vehicle spends at each detector. and The difference in the recorded timestamps.

3. The activity chain time allocation behavior modeling method considering travel costs according to claim 1, characterized in that, The utility function in step S2, which comprehensively considers the inherent utility and travel disutility of each activity, is decomposed into the following two product terms: and ; in, Indicates activity Negative effects of travel This represents the saturation parameter; Indicates time allocation parameters, The parameter representing the impact of the activity type. The parameter representing the impact of the activity duration, Indicate participation in the activity Duration Indicates arrival at the event The mathematical expression for the influence of location-related travel attributes on individual preferences is as follows: ； In the formula, Indicates travel distance. Indicates the number of traffic lights along the route. , They are respectively , The weight.

4. The activity chain time allocation behavior modeling method considering travel costs according to claim 1, characterized in that, The joint decision-making model for activity selection and time allocation established in step S3 is as follows: ； ； ； ； ； The joint decision-making model describes the decision-making process of individuals maximizing stochastic utility under time conservation constraints, where the time budget constraint is: ； in, This represents the total amount of time an individual can control. Indicate participation in the activity Duration Indicates travel time. This indicates the number of potential activity locations between two adjacent AVI detectors.

5. The activity chain time allocation behavior modeling method considering travel costs according to claim 1, characterized in that, In step S4, the joint decision model is solved, and the solution results are as follows: ； In the above solution results, This indicates the number of potential activity locations between two adjacent AVI detectors. Indicates the sample vehicles in The number of activities you choose to participate in during the activity. , Indicate participation in the activity Duration ; in, The intermediate variable is represented by the following expression: ； Indicates the number of participants The deterministic utility of an activity is determined by both the activity-related attributes and the travel attributes required to reach the activity location, as expressed below: ； in, This represents the saturation parameter; Indicates time allocation parameters, The parameter representing the impact of the activity type. The parameter representing the impact of the activity duration.

Images