Quantitative analysis and modeling of ride sharing behavior based on internet of vehicles

3.1 Influence of gender, age group, education, occupation background, private car, and income

The results of carpooling willingness corresponding to different genders are shown in Figure 1, a very small difference can be observed in the enthusiasm of males and females to participate in carpooling, and both are above 66%. It suggests that gender is not the key factor affecting the willingness to ride together.

Figure 1

The impact of gender on willingness of ride sharing.

The results of carpooling willingness corresponding to different age groups are shown in Figure 2. It is observed that travelers aged 18–40 have a higher enthusiasm for participating in carpooling, and all are above 60%. The proportion of travelers aged 41–60 who are unwilling to participate in carpooling is relatively high, with an average proportion of 63% unwilling to participate. It suggests that age is a key factor affecting the willingness to ride together.

Figure 2

The influence of age group on willingness of ride sharing behavior.

The results of the willingness to ride corresponding to the education level of travelers are shown in Figure 3. It is observed that the higher the level of education, the higher the enthusiasm for participating in carpooling. It indicates that education level is an important factor affecting the willingness to ride together.

Figure 3

The influence of education background on willingness of ride sharing behavior.

The results of the willingness of travelers to participate in carpooling corresponding to their professions are shown in Figure 4. It can be seen that different professions have a relatively small impact on whether they are willing to participate in carpooling. Except for merchants, the proportion of travelers from other professions who are willing to participate in carpooling is over 60%. From this, it can be seen that occupation is not the key factor affecting the willingness to ride together.

Figure 4

The influence of occupation background on willingness of ride sharing behavior.

The results of whether travelers have a private car corresponding to their willingness to share a ride are shown in Figure 5. It suggests that travelers without a private car are more inclined to share a ride than those with a private car. But regardless of whether they have a private car or not, the proportion of people willing to participate in carpooling is much higher than the proportion of people unwilling to participate in carpooling. It can be seen that the presence or absence of a private car is not a key factor affecting the willingness to travel together.

Figure 5

The impact of private car on willingness of ride sharing.

The impact of income on passenger carpooling willingness is shown in Figure 6, which indicates that low and middle-income travelers are more willing to choose carpooling, while high-income travelers have relatively lower enthusiasm for participating in carpooling. It can be seen that income is one of the main factors affecting the willingness to share rides.

Figure 6

The influence of income on willingness of ride sharing behavior.

The population willing to participate in carpooling has obvious group characteristics in terms of age, education level, income, etc. Young people are more receptive to new things and tend to choose the emerging mode of carpooling, while the elderly prefer traditional modes of transportation such as buses and bicycles.

Travelers with significant differences in economic conditions and education levels, due to their decision-making being influenced by economic affordability and inherent consumption concepts, have different choices of travel modes. Those with lower incomes are more inclined to choose private car sharing for cost savings, while those with higher incomes are more willing to choose private cars for privacy and comfort considerations. People with higher levels of education are more inclined towards environmentally friendly lifestyles and are more willing to try private car sharing for travel.

3.2 Factors leading to carpooling willingness

The factors leading to carpooling willingness were analyzed and are listed in Table 1. It is observed that the significant reason is “Save transportation costs,” which means that travel cost is an important factor for ride sharing.

The response percentage and case percentage of the influencing factors of willingness are also shown in Figure 7, and it is found that saving transportation costs is the main influencing factor. In addition, it is beneficial to alleviate traffic congestion and reduce parking pressure, which is also an important factor affecting carpooling. “Comfortable and convenient” is also a factor that leads to the willingness of ride sharing, which accounts for 12.8%.

Figure 7

Influencing factors of willingness: (a) response percentage and (b) case percentage.

3.3 Factors leading to unwillingness

The analysis of the factors that lead to unwillingness of carpooling was conducted and the results are listed in Table 2 and shown in Figure 8, with response percentage and case percentage.

Figure 8

Influencing factors of unwillingness: (a) response percentage and (b) case percentage.

Comfort, inconsistent starting points requiring detours, and long waiting times for carpooling are the main influencing factors for choosing non-carpooling options. Difficulty in coordinating among passengers and unclear sharing of risks and responsibilities are important factors for non-carpooling options, each accounting for around 13%. Travelers believe from a legal and safety perspective that unclear risk and responsibility sharing, as well as unsafe private car sharing, are the influencing factors for choosing not to ride. Cost disputes that are prone to occur during the ride sharing process are not the main influencing factors for this type of travelers, which accounts only for around 6.6%.

3.4 Factors encouraging ride sharing

For travelers who do not choose private car sharing for travel, various incentive measures would have an impact on their willingness to share. The results of these influences are shown in Table 3 and Figure 9. It suggests that the government’s formulation of corresponding safety measures and providing ride sharing subsidies will increase the proportion of passengers sharing the same ride. In addition, establishing a reasonable ride sharing fee system and setting up dedicated lanes for ride sharing will also increase the proportion of passengers sharing the same ride.

Figure 9

Influencing factors for encouraging ride sharing: (a) response percentage and (b) case percentage.

3.5 Key influencing factors for ride sharing

During the ride sharing process, due to the inconsistency between time and space, additional waiting and detour time will be generated, which is called additional time. When travelers travel for commuting, they generally choose a travel mode that is highly reliable in time and cost-effective. The impact of increasing travel time on carpooling behavior is shown in Figure 10a. From Figure 10a, it can be seen that travelers who accept less than 10% account for 60%, while travelers who accept 10–30% account for 37%. From this, it can be seen that the less additional time, the more people choose to ride together.

Figure 10

Key influencing factors for ride sharing: (a) travel time, (b) travel expense, and (c) number of passengers.

The fee standard refers to the total cost incurred by travelers in choosing a certain mode of transportation to achieve their free travel goals. Generally speaking, the higher the total cost of a certain mode of transportation, the less likely it is for travelers to choose it. According to Figure 10b, the proportion of travelers who accept travel expenses less than or equal to half of the normal travel expenses is 98%.

Due to the limited space and capacity of private cars, the number of passengers sharing can affect comfort. According to Figure 10c, it can be seen that the proportion of travelers who accept a carpool of two people is the highest, at 50%. The more people a private car can share, the higher the time and cost, and lower the quality and comfort of sharing. This can affect people’s willingness to share private cars.

The inconsistency between time and space is directly related to the proportion of carpooling trips. Therefore, in order to reduce the additional waiting and detour time during the carpooling process, it is necessary to encourage travelers to carpool through government policies and priority management measures. For example, the government can stimulate more travelers to choose private car carpooling trips by providing carpooling subsidies, setting up dedicated carpooling lanes, and setting up dedicated carpooling parking spaces.

3.6 Modeling and quantitative analysis private car sharing behavior

Based on the above investigation and analysis, this part will use game-theory to establish a dynamic game model for private car sharing behavior, and further analyze the impact of sharing related incentive policies on sharing behavior.

To develop of a Game-theory-based model, assuming that the proportion of private car owners choosing carpooling is x, the proportion of choosing non-carpooling is (1 – x); assuming the probability of the government adopting incentive policies is y, the probability of not adopting incentive policies is (1 – y). The resulting multiplication benefit matrix is shown in Table 4.

Here U _c is the cost savings and priority use of road rights obtained by travelers choosing private car sharing for travel. U _p is the flexibility, comfort, and path freedom that travelers can gain when choosing a private car to travel alone. T _c is when the government encourages carpooling, subsidies can be obtained from government departments for choosing carpooling. H _g is when the government encourages carpooling, the benefits brought to the government by alleviating traffic congestion and reducing exhaust emissions. N(x) is when the government does not encourage carpooling, the time value lost by carpoolers, as the number of carpoolers increases due to the lack of priority of right of way, is calculated using the following formula: N(x) = x · N, where N is the time loss cost caused by each additional carpooler.

C _g is the cost borne by the government in formulating policies related to carpooling. S _g is the losses borne by the government, such as pollution and congestion due to travelers abandoning private car sharing for travel.

Based on the above, it can be concluded that the expected benefits E ₁ for travelers choosing carpooling, E ₂ for choosing non-carpooling, and the average expected benefits Ē are:

The expected benefits of government support for carpooling I ₁, expected benefits of non-support for carpooling I ₂, and the average expected benefits Ī can be calculated separately using the following formulas:

Based on Eqs. (3) and (6), the dynamic equations for travelers and the government can be expressed as follows:

For travelers:

For government:

3.7 Analysis of equilibrium points in game model

The first step is to solve the equilibrium point. To make the mixed service mode selection, game model has an evolutionarily stable-strategy, it is necessary to satisfy both Eqs. (9) and (10) simultaneously.

From Eq. (9), the following transformation can be obtained:

By solving Eq. (11), it can be obtained that x = 0 , x = 1 , x = U c − U p + y T c ( 1 − y ) N . Assuming that f ( x ) = x ( 1 − x ) [ U c − U p + y T c − x ( 1 − y ) N ] , then there is

Assuming U c + T c U p > 1 , i.e., the total utility obtained by travelers choosing private car sharing for travel plus government subsidies exceeds the total utility obtained by private car traveling alone.

1. When 0 < y < U p − U c T c , then f ′ ( 0 ) < 0 , f ′ ( 1 ) > 0 , x ∗ = 0 can be obtained as evolutionary stability strategy, which means that when the effectiveness of government policies supporting carpooling is less than a certain threshold, travelers will ultimately abandon carpooling and choose to travel alone in private cars.

2. When U p − U c T c < y < U p − U c + N T c + N , f ′ ( 0 ) > 0 , f ′ ( 1 ) > 0 , f ′ U c − U p + y T c ( 1 − y ) N < 0 , x ∗ = U c − U p + y T c ( 1 − y ) N can be achieved as evolutionary stability strategy, where the policy validity of government support for carpooling is within a certain range, the proportion of people choosing private car carpooling will eventually stabilize within the range of [0, 1].

3. When U p − U c + N T c + N < y < 1 , there is f ′ ( 0 ) > 0 , f ′ ( 1 ) < 0 , x ∗ = 1 as evolutionary stability strategy, where the policy validity of the government’s support for carpooling exceeds a certain threshold, travelers will ultimately choose carpooling, which theoretically proves the necessity of the government’s formulation of carpooling incentive policies. At this point, the following is satisfied:

By solving Eq. (13), it can be obtained that y = 0 or 1. When x = C _g/H _g, all the y are in stable state; however, when x ≠ C _g/H _g, y = 0 and y = 1 are two stable points. Assuming f(y) = y(1 – y)(xH _g – C _g), then f’(y) = (1 – 2y)(xH _g – C _g), thus when 0 < x < C _g/H _g, f′(0) < 0, f′(1) > 0, y ^* = 0 as evolutionary stability strategy, the proportion of private car owners choosing carpooling is less than a certain threshold, the government does not need to take any incentive measures.

When C g / H g < x < 1 , f ′ ( 0 ) > 0 , f ′ ( 1 ) < 0 , then y ∗ = 1 as evolutionary stability strategy, when the proportion of private car owners choosing carpooling exceeds a certain threshold, the government needs to formulate incentive policies related to private car carpooling to improve the efficiency of carpooling.

Thus, the local equilibrium points of the dynamic system composed of the above evolutionary game model are (0, 0), (0, 1), (1, 0), (1, 1), U c − U p N , 0 , C g H g , ( U p − U c ) H g + C g N H g T c + C g N . By analyzing the stability of the Jacobian matrix of the system, the stability characteristics of the system equilibrium point can be obtained. By replicating the dynamic equation and differentiating x and y separately, the Jacobian matrix of the mixed service mode selection game model can be obtained:

The determinant (Det) and trace (Tr) of the matrix are as follows:

The Jacobian matrix results of the mixed service mode selection game model are shown in Table 5.

3.8 Analysis of local stability at equilibrium points

According to the values in Table 5, to determine the symbols of Det (J) and Tr (J) corresponding to each equilibrium point, it is necessary to divide the relationship between parameters into different scenarios for judgment. It can be divided into two categories as a whole, namely, the utility obtained by private car carpooling is smaller than that obtained by private car solo travel, and the utility obtained by private car carpooling is greater than that obtained by private car solo travel.

In the scenario where the utility obtained from private car carpooling is less than that obtained from private car solo travel, there are several situations.

(1) When U c U p < 1 , H g C g < 1 , the local stability of the equilibrium point is shown in Table 6.

(2) When U c U p < 1 , H g C g > 1 , due to conflicting conditions, this condition is not discussed here.

In the scenario where the utility of private car sharing is greater than that of private car traveling alone, there are several situations.

When 0 < U c − U p N < 1 , H g C g < 1 , the local stability of the equilibrium point is shown in Table 7.

However, when 0 < U c − U p N < 1 , H g C g > 1 , U c − U p N < C g H g , the local stability of the equilibrium point is listed in Table 8.

It is observed that the final stable points of the system are U c − U p N , 0 and (1, 1), and the initial state of the system will have a decisive impact on the final convergence stability strategy.

When 0 < U c − U p N < 1 , H g C g > 1 , U c − U p N > C g H g , the corresponding local stability of the equilibrium point is listed in Table 9.

When U c − U p N > 1 , H g C g > 1 , the corresponding local stability of the equilibrium point is listed in Table 10.

When U c − U p N > 1 , H g C g < 1 , due to conflicting conditions, this condition is not discussed here. In summary, the stability points of the system include the following three points: (0, 0), U c − U p N , 0 , (1, 1). The equilibrium conditions are presented in Table 11.

I When the evolutionary conditions meet U c U p < 1 , H g C g < 1 , (0, 0) is the equilibrium point of the system, i.e., when the utility of choosing private car sharing for travel is less than that of traveling alone, and the benefits obtained from the government’s implementation of sharing related incentive policies are less than the cost of policy formulation, travelers will ultimately choose to travel alone, and the government will not adopt any incentive policies. The root cause of this situation is that the transportation system operates relatively smoothly, and the benefits of private car non ride travel are higher, so more people tend to choose non ride travel.

II When the evolutionary conditions meet 0 < U c − U p N < 1 , H g C g < 1 , or 0 < U c − U p N < 1 , H g C g > 1 , U c − U p N < C g H g , the equilibrium point of the system is U c − U p N , 0 . This situation generally occurs in the early stages of private car sharing, where people are not yet very familiar with this new mode of transportation. Only some young people are willing to try sharing transportation, and government management departments are also holding a wait-and-see attitude toward the development of sharing transportation. The feasibility of relevant incentive policies is currently in the research stage.

III When the evolutionary conditions meet U c − U p N > 1 , H g C g > 1 , (1, 1) is the equilibrium point of the system. In this situation, the operation of the transportation system is already in a congested state, and the utility of choosing private car sharing for travel is greater than that of traveling alone. Moreover, the benefits obtained by the government from sharing for travel are also greater than the cost of formulating policies. Therefore, more travelers ultimately choose private car sharing for travel, and the government will develop incentive policies and priority management measures related to sharing for travel driven by benefits, in order to fully improve the efficiency of sharing for travel, attract more travelers to share for travel, and thus form a virtuous cycle.