Evaluation of Different Work Zone Road-Occupation Schemes for Monorail Construction

Abstract

Due to the increasing demand for transportation, road renovation is inevitable, and the emergence of work zones has become the most common problem in traffic control. There are many research results on the impact of work zones on traffic operation, but most of them focus on various capacity theoretical models or traffic flow control strategies and are mostly concentrated in highway work areas and urban road subway construction work areas. The location of these work zones is often unable to be changed and is mostly considered reasonable. At present, there is still a gap in selecting and evaluating the location of the work zone. Therefore, this article studied a bidirectional six-lane intersection in Xi’an and investigated the situation of traffic flow distortion caused by the setting of work zones at the intersection. Two different positions and forms of work zones were designed for evaluation and analysis. Firstly, we used VISSIM (2022 student) to simulate the original and proposed work zone road-occupation schemes and analyzed six indicators. Finally, in order to determine the optimal work zone locations under different traffic situations, we applied the entropy weight method (EWM) to assign weights to multiple indicators, thereby achieving a comprehensive assessment of various schemes. The results showed that the highest improvement level among the six indicators was 50.2%, and different schemes adapted to different traffic situations. However, occupying two lanes of the median opposite the exit lane performed better under high traffic volume. Occupying the median and lanes on both sides of the median is suitable for low traffic volume. Occupying two lanes of the entrance lane and using the opposite lane as the left-turn entrance lane is suitable for situations with high traffic volume in the opposite lane.

1. Introduction

The rapid development of China’s transportation industry has led to a swift increase in the number of vehicles in urban areas, and the traffic pressure in cities is increasing gradually. In order to alleviate the congestion and air pollution caused by the imbalance between urban transportation supply and demand [1], in addition to accelerating the construction of subways and urban expressways, China has been vigorously developing monorail transportation in the past two years, constructing tracks on the central dividing strip or narrow streets of the road. This type of rail transit does not occupy the road surface separately and belongs to a medium-capacity urban rail transit system with a capacity similar to that of a subway system [2]. Although the monorail does not occupy the road surface, it needs to occupy the urban ground to build the foundation or place construction tools during the construction process. During the construction process of a monorail, 2–3 lanes are generally required, with only the median strip of the road being used to build the foundation and other places being used to place construction equipment. Whether the work zone is situated on the road or near the intersection, its existence will inevitably have an impact on the traffic flow passing by [3,4]. Due to the long construction length, the monorail work zone often occupies the entire lane of two adjacent intersections and seriously distorts the traffic flow of vehicles at the entire intersection [5], leading to a decrease in the traffic capacity of the intersection [6,7]. And intersections are already the bottleneck of urban road traffic. Setting up work near the bottleneck of urban roads will inevitably further exacerbate traffic congestion [8], even if it would endanger traffic safety [9].

Over the years, in an effort to alleviate the effects of construction work zones on traffic efficiency, researchers and policymakers in the field of transportation have been exploring various strategies. Some scholars have evaluated the impact of road construction through dynamic traffic allocation and compared alternative construction schedule plans in order to choose the optimal plan to reduce travel delays and improve road capacity [10,11]. Several studies have analyzed the impact of construction work zone length on road capacity [12,13,14]. In addition, scholars have found that a work zone located at the intersection causes deviation of the road, which in turn leads to deviation of the main traffic flow, and have proposed two road improvement plans for this purpose [15]. Due to the limited location selection of the construction work zone studied above, the main idea is to improve the form of the construction work zone to ensure smooth traffic flow and increase road traffic capacity. When constructing a monorail, the location of the ground construction work zone is relatively flexible, and different middle lanes can be chosen as the work zone (including the median strip) according to the actual situation. In most practical situations, the setting of work zones is often selected based on the actual form of land required for maintenance or construction. This type of work zone is objective, cannot be selected, and has a relatively short existence time with little impact on traffic. However, the construction of monorail transportation takes longer than ordinary construction operations, usually taking several months or even years. Reducing the impact of monorail work zones on intersections has become an urgent issue to be addressed.

Therefore, this paper primarily investigates the influence of work zone layout on intersections during the construction of monorail systems. Research on the effects of work zones predominantly falls into two major categories. The first category pertains to the work zones on highways, encompassing aspects such as traffic capacity [16], work area impact area [17], and traffic flow prediction models [18]. The second category focuses on urban work zones, including studies related to urban travel origin–destination (OD) matrices [19], alternate route selection [20], and traffic organization management [21]. Existing studies have notably overlooked research into the layout and location selection of work zones, and there has been limited investigation into the corresponding rationality of work zone layout and road geometry design. Furthermore, the establishment of work zones at intersections has demonstrated substantial adverse effects on sustainable traffic flow.

This study selects the monorail work zone at an intersection in Xi’an as the research subject. Two monorail work zone layouts and location schemes were proposed, and corresponding intersection road layout schemes were proposed based on the actual data collection location of setting up the work zone at intersections. The entropy weighted method (EWM) is an objective, comprehensive evaluation method. Based on on-site survey data, we used VISSIM simulation [22,23,24] and the EWM [25,26,27] to calculate and evaluate the three schemes and selected the best scheme under different traffic volume combinations; we utilized VISSIM software to simulate various schemes. Ultimately, using the EWM, we conducted a comprehensive scoring of different schemes to determine the optimal construction organization method under varying traffic volumes.

The remaining sections of this article are outlined as follows. Section 2 provides a literature review of relevant studies. Section 3 presents the problem statement and describes the collection of case data. In Section 4, we modeled, calibrated, and analyzed intersections with work zones occupying different lane positions. In Section 5, we conducted a performance improvement analysis of the three proposed schemes for different traffic volumes. In Section 6, we utilized the entropy weighted method (EWM) to comprehensively evaluate the three schemes, calculate and describe the final selected scheme, and draw conclusions. Section 7 provides a summary of the paper.

2. Literature Review

Through a literature review, it has been found that research on the impact of construction work zones on traffic can be broadly categorized into two main types. The first type involves analyzing and summarizing the effects of construction work zone configurations on nearby traffic safety and variations in traffic variables. The second type employs advanced instruments to collect actual data, including the geometric dimensions of work zones and traffic data, for the purpose of traffic flow modeling.

Raju, N. [28] studied the lateral driving behavior of drivers in work zone environments, compared and analyzed the macroscopic traffic characteristics with and without construction work zones, and found that the vehicle speed decreased from 70 km/h to 50 km/h, and the efficiency of each lane capacity decreased. At the same time, based on vehicle trajectory data, it was found that the normal lane capacity value was higher than the road with a construction work zone. Alshabibi, N. M. collected [21] traffic data from all work zones in New York during and after their removal, revealing that the saturation progress during the work area was relatively modest compared to the period following removal. This implies an increase in the saturation flow for each channel during the work area. When there are people or heavy equipment in the work area, drivers often adhere to the speed limit in the work area [29]. Kamyab, M. [30] used machine learning to predict the influencing factors of future work areas and found that long-term speed changes are important factors in predicting the impact of work areas on traffic. Pesti, G. [10] developed a method for evaluating the impact of road construction work areas, (a) predicting the network level impact of road construction projects, (b) identifying key sections and corridors with the most severe expected construction impact, and (c) comparing alternative construction plans and schedules. Dynamic traffic allocation forms the basis for evaluating the regional impact of road construction and comparing alternative construction schedule plans. Venter, L. [31] studied the impact of lane occupancy in work areas on road capacity, determined queue positions and the maximum waiting time of vehicles based on changes in traffic volume, and calculated the length of construction work areas suitable for occupying half of the lane. Scholars have also used VISSIM (2022 student) simulation to study the impact of subway track construction work areas on the transportation environment from the perspective of environmental protection [17] and estimated the economic losses caused. The study by Liu, Y. [32] is based on microvehicle operation, establishing a framework for the relationship between maintenance work areas and carbon dioxide emissions. Preventive maintenance of vehicles in the work area is the main source of carbon dioxide emissions, especially the variable speed and queuing vehicles near the work zone.

Fang, S. [33] conducted a study on the impact of low-speed mobile construction zones on urban roads. The findings revealed that when traffic volume reached 2000 pcu/h, each lane was significantly affected, mainly due to lane-changing behavior in the rear lane of the low-speed mobile construction zone. At the same time, as the speed of the mobile construction zone increased, the average traffic risk decreased. The scholar proposed a microscopic model of traffic flow in a straddle work zone to capture the characteristics of no lane division and irregular boundaries that intersect with the straddle work zone, reflecting the effect of traffic flow from two opposite directions of the intersection road [11]. Finally, the entropy method was used for evaluation, and the results showed that the distance from the lower edge of the work zone to the centerline and the proportion of large vehicles occupying the work zone exerted the most significant influence on the signalized intersection. Weng, J. [14] proposed three workspace speed flow and capacity models based on the measured number of workspaces, geometric alignment, workspace restricted speed, and workspace length at different locations. These models can examine the impact of workspace configuration factors on the workspace speed flow relationship and capacity. The results show that traffic speed, flow, and workspace capacity increase with the increase in restricted speed. The increase in the number of jobs has a greater negative impact on traffic flow than the increase in geometric design. In 2019, the Bayesian method was used to calibrate the model and establish a work zone capacity prediction model that follows a lognormal distribution. Feng Deheiden, N. analyzed a large number of short-term work zones with temporarily closed lanes and long-term work zones without reducing the number of lanes. Based on the differences between the two, the former used deterministic and stochastic capacity estimation methods, while the latter only determined capacity for congested flow after failure.

Due to the significant impact of work zone settings on traffic safety, many scholars have established models to predict conflicts and collisions near the work zone, including methods such as the artificial neural network by Cheng, Y. [34] and the minimum median quadratic linear regression model. Cheng, Y. [35] analyzed the safety impact methods of using signal lane control strategies at work area merging points and found that signal control devices can reduce lane-change conflicts at work area merging points. Hou G. [36] proposed a comprehensive framework for evaluating traffic safety in work areas under adverse driving environments, introducing a new indicator that integrates the risks of years of collisions and single-vehicle collisions to assess traffic safety risks in work areas. Ghasemzadeh, A. [37] developed an ordered probability model that considers different spatial, temporal, and environmental conditions as factors for the severity of work zone collapse. The results indicate that these are decisive factors in increasing the severity of work zone crashes.

Through literature analysis, it can be seen that existing research results mostly focus on management strategies, strategies for controlling traffic flow, the establishment of various models, and the impact of work areas on roads.

3. Problem Statement and Data Collection

3.1. Problem Statement

In typical scenarios, urban road expansion construction often involves the occupation of a particular lane due to the need for maintenance or road widening. During construction, it becomes necessary to temporarily occupy the target lane.

The rapid increase in the number of motor vehicles in Chinese cities has led to frequent and severe congestion on urban main roads and intersections during morning and evening peak hours. In order to alleviate traffic pressure, more and more places are renovating and expanding urban roads. In addition to widening or designing the alignment of urban roads, cloud rail (suspension rail system) construction has emerged in China in the past two years. The tracks in the suspension rail system can be built on the central median of the road or narrow streets without occupying the road surface separately. It belongs to the medium-volume urban rail transit system with a capacity close to the subway system. During the construction of cloud tracks, it is often necessary to occupy the middle part of urban roads (lanes or central separation strips) for construction equipment stacking or cloud track foundation construction, usually occupying 2–3 lanes during construction. Choosing which lane to occupy in the middle under different traffic conditions has less impact on intersection traffic, which has become an urgent issue for urban reconstruction and expansion projects in China at present. Therefore, this article takes the construction of a monorail at the intersection of a main road in a certain city of Xi’an as a case study. The objective is to investigate the effects of occupying different lanes in the middle during the construction of the monorail on the traffic at the intersection. Additionally, the article offers a calculation method to assist in determining suitable solutions. The intersection is the main road of two urban roads, Zhangba Fourth Road and Jinye Second Road (as shown in Figure 1). The initial lane configuration at this intersection consists of six lanes. There is an auxiliary lane on each side, which is often used as a dedicated lane for buses unless motor vehicles are driving. The speed limit for the north–south main road within the construction operation area is 40 km/h, and the speed limit for the east–west main road is 60 km/h. During the construction period, the innermost lanes of the south entrance and north exit of Zhangba Fourth Road at the intersection were occupied by the construction area. To ensure a balanced number of lanes and smooth traffic, the auxiliary lanes on both sides were used as motor vehicle lanes.

3.2. Data Collection

In VISSIM (2022 student), the calibration and microscopic simulation of models typically necessitate actual traffic data measurements. (Figure 2—actual diagram of the intersection). According to Auto Navi Traffic statistics, the peak traffic congestion on Zhangba Fourth Road in Xi’an is between 7:30 and 8:30, as well as between 17:30 and 18:30. Therefore, on-site data collection is conducted during peak and undervalued traffic flow times. Throughout the data collection period, favorable weather and road conditions prevailed, with no reported traffic accidents (

Table 1. Traffic volume statistics during one peak hour (7:30 a.m. to 8:30 a.m.) on 17 June 2023.

Direction	Steering	Car	Truck	Bus
E1	Straight	297	4	8
	Left	78
	Right	45
E2	Straight	315	1	8
	Left	39
	Right	115		1
W	Straight	470	15	8
	Left	112	1	6
	Right	57
S	Straight	414	24	1
	Left	63	1
	Right	178	3	1
N	Straight	280	20	6
	Left	66	1	1
	Right	92		1

E1 represents the motor lane of eastbound road; E2 represents the nonmaneuverable lane of eastbound road; W represents the westbound road; S represents the southbound road; N represents the northbound road.

).

The traffic volume collection method in this article employs an unmanned aerial vehicle (UAV), which statistically classifies vehicles passing through intersections. The collected data includes the following:

• Traffic volume in all directions at the intersection of Zhangba Fourth Road and Jinye Second Road.
• The proportion and the type of left and right turns at each entrance.
• The geometric shape and width of the occupied road during intersection construction.

After data processing, the peak traffic volume hours at the intersection is 3420 veh/h, and the collected traffic data has the following characteristics:

• There is a certain disparity in traffic volume between the two main roads at signalized intersections, with traffic flow in the east–west direction significantly greater than that in the north–south direction.
• The left-turn rate of traffic flow at the east entrance of the signalized intersection is 10%, and the left-turn rate in other directions is 18%.
• Due to construction occupation, the east entrance occupies a nonmotorized lane, and the traffic volume of this lane is basically the same as the normal two-lane traffic volume.

4. VISSIM Simulation

4.1. Establishing the Model

VISSIM (2022 student) is a microlevel simulation modeling tool based on time intervals and driving behavior used for traffic modeling of urban and public transportation operations. Therefore, this study uses VISSIM (2022 student) to simulate and evaluate solutions. Firstly, AutoCAD (2023) is used to draw existing actual solutions and two improved solutions, and they are imported into VISSIM (2022 student) as a base map. Secondly, to ensure the accuracy of the simulation model, the traffic parameters such as vehicle speed, traffic volume, geometric parameters (i.e., number of lanes, lane width, and workspace position), vehicle path, and traffic flow ratio in the model should be consistent with the actual situation and set one by one. In addition, to meet the actual traffic driving situation and conditions, conflict zones and signal control light phase parameters should also be set at intersections and east entrance roads.

In order to study the impact of different occupation positions on traffic in the construction work zone of intersections, this paper considers three different construction occupation positions based on measured data: the innermost two lanes in the opposite direction, the innermost two lanes in the entrance road, and the innermost lanes in the entrance and exit roads. The traffic organization forms for the three schemes are shown in Figure 3. Figure 3a is the actual scheme used at the intersection, which occupies the innermost lane of the south entrance and exit while using the outermost auxiliary lane as a right-turn lane (Scheme 1). Figure 3b involves occupying the innermost two lanes of the south entrance while also occupying the opposite lane as a left-turn lane and the auxiliary lane as a right-turn lane (Scheme 2). Figure 3c involves occupying the two lanes from the south exit to the innermost side. To maintain lane balance, the left-turn lane of the entrance lane is used as the opposite straight lane (Scheme 3).

VISSIM (2022 student) was used to simulate three road occupancy schemes. To achieve more precise simulation results, based on on-site investigations and suggestions from local transportation departments, some lane-changing parameters have been fixed to ensure that the simulated driving behavior aligns with that of Chinese drivers, as shown in

Table 2. Simulation of driving parameter correction.

Maximum Deceleration (m/s2)	Accepted Deceleration (m/s2)	Safety Distance Reduction Factor	Coordinate Maximum Deceleration (m/s2)
−4	−2	0.5	−4

:

4.2. Calibration of the Simulation Model

During the VISSIM (2022 student) simulation, in order to ensure the accuracy of the results, it is necessary to calibrate the model parameters based on on-site data. The collected parameters include traffic data, signal period, road geometry, etc. The calibration program is based on previous research [23,38].

The establishment of traffic capacity is an important parameter for path selection in VISSIM (2022 student) simulation, which can reflect various attributes in the model; therefore, accurate road capacity is required. The capacity calculation formula is as follows:

$$C=\frac{3600}{\overline{h_t}}$$

(1)

where C denotes the ideal capacity (veh/h), and h_t is the average minimum headway (s).

Various collected traffic data were input into the VISSIM (2022 student) simulation model. If the output capacity of the VISSIM (2022 student) simulation model is close to the actual collected capacity, it can prove that the VISSIM (2022 student) simulation model is reasonable. The error between simulated data and collected data is calibrated using the mean absolute percent error (MAPE) indicator. The MAPE indicator represents the average absolute percentage error. When the MAPE is less than 15%, it indicates that the model’s accuracy meets the requirements [39,40].

The MAPE formula is as follows:

$$MAPE=\frac{{\displaystyle \sum _{i=1}^n{C}_v^i}-{\displaystyle \sum _{i=1}^n{C}_f^i}}{{\displaystyle \sum _{i=1}^n{C}_f^i}}$$

(2)

where i denotes the traffic flow, n denotes the total number of traffic flows, C_vⁱ denotes the simulated traffic volume (veh/h), and C_fⁱ denotes the measured traffic volume (veh/h).

Table 3. MAPE calculation direction.

	Steering	Investigated Capacity (veh/h)	Simulated Capacity (veh/h)	Single MAPE (%)	Mean MAPE (%)
E1	straight	387	55	−2.3%	7.31%
	Left	97	11	18.35%
	right	54	6	20%
E2	straight	408	58	−2.4%
	Left	60	9	−8%
	right	132	19	−3.6%
W	straight	604	78	7%
	Left	151	17	18.9%
	right	84	11	5.7%
S	straight	549	64	16.1%
	Left	86	12	0.5%
	right	223	33	−6.5%
N	straight	380	45	14.7%
	Left	88	13	−6.4%
	right	117	10	38.5%

E1 represents the motor lane of eastbound road; E2 represents the nonmaneuverable lane of eastbound road; W represents the westbound road; S represents the southbound road; N represents the northbound road.

shows the MAPE results of traffic flow in each direction of the intersection, and the average value can be obtained to obtain the MAPE results of the entire intersection. The MAPE calculation results reflect a total error of 7.31% between the VISSIM (2022 student) simulation model and the measured traffic volume. This indicates that the VISSIM (2022 student) error is within an acceptable range and its accuracy meets the standard. At the same time, the MAPE result of the right-turn direction at the northbound road reaches 38.5%, with a large error result. However, this article mainly studies the impact of the eastbound and westbound roads on the intersection and the changes in various indicators. Moreover, the right-turn vehicles on the northbound road do not affect the driving situation on the eastbound road, and the MAPE of the entire intersection also meets the standards. Therefore, it can be considered that the VISSIM (2022 student) simulation model in this article meets the requirements.

4.3. Simulation Results

Due to the different forms of road occupation at intersections affecting the overall traffic operation efficiency, the traffic operation efficiency of intersections with different schemes was evaluated. Usually, the most commonly used indicators for evaluating the operational efficiency of intersections are delay and stop frequency. In addition to both, this study also takes the queue length and the number of vehicles as evaluation indicators. Based on environmental protection considerations, the emission-related parameter CO emissions in VISSIM are considered as one of the evaluation indicators.

From the indicator results of different schemes in

Table 4. Simulation results of three schemes.

Indicator	Scheme 1	Scheme 2	Scheme 3
queue length (m)	34.03	28.24	32.61
number of vehicles (veh)	443	460	455
Delay (s)	33.33	34.37	32.89
Stops Frequency	0.81	0.85	0.80
CO emissions (grams)	702.06	654.46	669.44

, it can be seen that there is not much difference in the indicators of the three schemes under the current actual traffic volume situation. However, except for the delay indicator of Scheme 1, which is better than Schemes 2 and 3, all other indicators are not as good as the other two schemes. It can be seen that when the construction work zone occupies the inner lanes of both directions, it is more likely to cause congestion and reduce the traffic efficiency of intersections. On the other hand, there is not much difference in the results of each indicator between Schemes 1 and 3, but Scheme 3 outperforms Scheme 1. This may be due to the fact that although the construction occupation positions are different, the vehicle lanes at the east entrance of the two are the same, affecting only the traffic conditions of the inner lane at the west exit, thus affecting the traffic efficiency of the entire intersection. Scheme 2 has lower delay and stop frequency compared to the other two schemes, while other indicators are better than the other two schemes. Considering the performance of the three schemes under different combinations of traffic volumes, each exhibits its own set of strengths and weaknesses. Therefore, this article conducts a more comprehensive analysis of the three schemes to determine their performance accurately.

4.4. Safety Analysis

The complete simulation analysis includes traffic operation assessment and safety assessment. The safety assessment model used in this article is the SSAM (X86) model provided by the Federal Highway Administration, which is used to predict road safety conditions before traffic accidents. In addition, based on the angle between the two vehicle bodies at the time of the conflict, the conflict is divided into rear-end conflict (0°, −30°), lane-change conflict (30°, −85°), and intersection conflict (85°, −180°). The data collected during peak periods were brought into the following three scenarios, and the results are presented as follows.

According to the above judgment criteria, the conflict analysis results of the three schemes are shown in

Table 5. Statistics on types of traffic conflicts.

Scheme	Rear-End Collision	Crossing Conflict	Lane-Change Conflicts	Summary
Scheme 1	120	138	78	336
Scheme 2	105	161	82	348
Scheme 3	113	143	65	321

. Scheme 2 shows that when the left-turn lane at the east entrance occupies the target lane, the lane-change conflict and intersection conflict are higher than those of Schemes 1 and 2, while the rear-end conflict is smaller. The total number of conflicts in Scheme 3 is lower than that in Schemes 1 and 2.

5. Analysis of Behavior Improvement of Schemes

Traffic volume is an important factor affecting intersection capacity. The above analysis only focuses on the performance of each scheme under the measured traffic volume during peak hours and cannot reflect the traffic conditions across various traffic scenarios. The occupation of different lanes in the construction area affects not only the traffic efficiency of the entrance road but also the traffic operation of the opposite lane. Hence, it is essential to choose traffic volume in different entrance directions of urban intersections as sensitive factors to study the impact of construction occupation selection on intersections under diverse traffic conditions.

According to the collected traffic data and Highway Capacity Manual (HCM) [41], the maximum service traffic volume for dual-direction, six-lane urban roads under a 40 km/h speed limit is 2133 veh/h. In performance improvement analysis, based on the actual traffic capacity of different entrances, the traffic volume range of the east entrance is taken as 0.3–1.0 V/C, and the traffic volume range of the west entrance is 0.2–0.8 V/C. The traffic volume data for performance improvement analysis is shown in

Table 6. East and west traffic simulation.

Direction	Traffic Volume (veh/h)
E	711/948/1185/1422/1659/1896/2133
W	474/711/948/1185/1422/1659/1896

E represents the eastbound road; W represents the westbound road.

. In the performance improvement analysis, changes are made only to the traffic volume at the east and west entrances, which has the most significant impact on the operational efficiency of the intersection, and the traffic volume at the north and south entrances remains constant.

A total of six indicators, queue length, delay, number of vehicles, stopping frequency, and CO emissions from the VISSIM (2022 student) simulation results, as well as the number of conflicts obtained from the assessment model (SSAM X86) analysis, were selected for the performance improvement analysis.

In Figure 4 and Figure 5, the x, y, and z coordinates represent different traffic volumes at the west entrance, different traffic volumes at the east entrance, and the degree of performance improvement for each indicator. They, respectively, show the control situation of Schemes 2 and 3 compared to Scheme 1 at the intersection, evaluating the superiority of 48 traffic combinations in six indices. The positive indicator is better than the original scheme, while the negative indicator is inferior to the original scheme.

Figure 4a,b, respectively, show the improvement degree of queue length and vehicle quantity in Scheme 2 compared to Scheme 1. The improvement trend of both is consistent. When the traffic volume at the east entrance is small, Scheme 2 outperforms Scheme 1 at the intersection, with maximum improvement rates of 54.4% and 65.8%, respectively. As the traffic volume at the east entrance gradually increases, the improvement effect of Scheme 2 gradually decreases to be consistent with the effect of Scheme 1. On the other hand, the improvement rate decreases with the increase in traffic volume at the west entrance, but the change is relatively small.

Figure 4c illustrates the degree of improvement in CO emissions for Scheme 2, similar to the overall trend of queue length and vehicle quantity. The maximum improvement rate of Scheme 2 is 36.6%. When the east entrance traffic volume for Scheme 2 exceeds 1896 veh/h, the improvement degree in CO emissions compared to Scheme 1 noticeably decreases.

Figure 4d shows the improvement rate in terms of parking frequency for Scheme 2 compared to Scheme 1. When the east entrance traffic volume is low, Scheme 2 and Scheme 1 have similar operational performance. When the traffic volume at the east entrance is 711 veh/h, the improvement rate is the highest at 18.0%. As the traffic volume at the east entrance increases, the performance of Scheme 2 is noticeably inferior to that of Scheme 1, and the most unfavorable situation leads to a 50% increase in parking frequency.

As can be seen in Figure 4e, Scheme 2 has lower delays than Scheme 1 at lower traffic volumes, with the highest improvement rate increasing by 45.1%. The delays of Scheme 1 and Scheme 2 are similar when the east entrance traffic volume continues to increase up to a moderate level. When the traffic volume exceeds 1659 veh/h, Scheme 1 is superior to Scheme 2, and as the traffic volume increases, Scheme 1 gradually outperforms Scheme 2, with a maximum delay improvement rate of 17.3%.

Comparing the trend of conflict indicators between Schemes 1 and 2 (Figure 4f), in most traffic volume combinations, Scheme 1 is superior to Scheme 2, and the conflict indicator of Scheme 2 is 38.9% higher than that of Scheme 1 at the most unfavorable time. When the traffic volume at the east entrance is 711 veh/h and the traffic volume at the west entrance is less than 711 veh/h, Scheme 2 does not provide a significant degree of enhancement.

The performance improvement analysis results show that compared to occupying the innermost two lanes of the east entrance during the construction of the monorail (Scheme 2) and occupying the innermost single lane of the east entrance and west entrance (Scheme 1), Scheme 2 is suitable for situations with low traffic volume at the east entrance, while Scheme 1 is suitable for situations with high traffic volume. This indicates that occupying the opposite lane as the left-turn lane of the east entrance can ensure that the innermost predicted left-turn vehicles on the east entrance reduce queue length and ensure the travel time of left-turn vehicles. However, as the overall traffic volume of the east entrance increases, the negative impact of this plan on traffic is greater than its positive impact, especially when the traffic volume exceeds 1659 veh/h.

Compared with Scheme 1, Scheme 3′s queue length is superior to Scheme 1 in most traffic volume combinations, with a maximum increase of 37.1%. When the traffic volume at the eastbound is 711 veh/h, the improvement rate of Scheme 3 decreases. In the most unfavorable situation, Scheme 3 performs significantly worse than Scheme 1, with a queue length improvement rate of −14.4%.

The improvement degree of vehicle quantity and parking frequency between Schemes 3 and 1 is in Figure 5b,d. In all traffic volume combinations, Scheme 3 is superior to Scheme 1, with a maximum improvement degree of 29.7% and 36.1%, respectively. However, while the eastbound traffic volume is relatively low, there is little change compared to Schemes 3 and 1.

Figure 5c,e compare the CO emissions and delays between Schemes 1 and 3. Under the condition of low eastbound traffic volume, the improvement rate of Scheme 3 decreases with the increase in traffic volume at the west entrance. When the westbound traffic volume exceeds 1422 veh/h, Scheme 1 is significantly better than Scheme 3. Overall, Scheme 3 is superior to Scheme 1, with Scheme 3 achieving the highest improvement in CO emissions and delays of 20.4% and 28.1%, respectively.

Figure 5f compares the conflict situation between Schemes 1 and 3. As the eastbound traffic volume increases, Scheme 3 gradually outperforms Scheme 1, and the growth rate of conflict in Scheme 1 is significantly higher than in Scheme 3. Scheme 3 achieves a maximum improvement of 50.2% in high traffic conditions.

The performance improvement analysis results show that in most traffic combinations, the construction occupation of the two innermost lanes of the west exit (Scheme 3) is better than the occupation of each inner single lane of the east entrance and west exit (Scheme 1). However, in the case of low traffic volume at the east entrance but high overall traffic flow at the intersection, Scheme 1 is better than Scheme 3.

6. Analysis of the Results Based on EWM

6.1. Entropy Weight Method

The EWM is a comprehensive evaluation method that can be used for multiple objects and indicators. In the specific usage process, the EWM calculates the weight of each indicator based on the degree of variation of each indicator, uses information entropy to calculate the weight of each indicator, and then corrects the weight of each indicator through entropy weight to obtain more objective indicator weights, and finally performs scoring. The EWM is commonly used to comprehensively compare multiple solutions and calculate the optimal solution in complex situations, which can avoid the subjectivity and one-sidedness of evaluation results.

6.2. Evaluation of Schemes

Through performance improvement analysis, the performance of the suspension railway work zone at this signalized intersection, which occupies different lanes compared to the original scheme, was discussed under 49 traffic combinations. The results indicate that under different traffic combinations, every scheme emerges with different advantages in six indicators. Therefore, a method is to propose finding the optimal work zone solution for each group of traffic volume.

Firstly, the combination of simulation results for each indicator was converted into a matrix X of 49 rows and 6 columns under 49 traffic volume situations:

$$X_i=\left[D_{\mathrm{i},\mathrm{n}},{\mathrm{Q}}_{\mathrm{i},\mathrm{n}},S_{\mathrm{i},\mathrm{n}},V_{\mathrm{i},\mathrm{n}},E_{\mathrm{i},\mathrm{n}}.C_{\mathrm{i},\mathrm{n}}\right]$$

(3)

In matrix X₁,

$$D_{1,\mathrm{n}}=\left[D_{1,1},{\mathrm{D}}_{1,2},{\mathrm{D}}_{1,3},{\mathrm{D}}_{1,4},{\mathrm{D}}_{1,5},\dots ,{\mathrm{D}}_{1,49}\right]$$

(4)

Secondly, the three matrices, X₁, X₂, and X₃, were split and reorganized, and the differences between the solutions under 49 different traffic volume situations were compared. In each traffic volume combination, every simulation results of Scheme 1 to 3 into matrix An were merged:

$$A_{\mathrm{n}}=\left[\begin{array}{l}{X}_1\left(n,·\right)\\ X_2\left(n,·\right)\\ X_3\left(n,·\right)\end{array}\right]$$

(5)

Therefore, the obtained A_n is a matrix of three rows and six columns:

$$A_{\mathrm{n}}=\left[\begin{array}{l}{D}_{1,\mathrm{n}},{\mathrm{Q}}_{1,\mathrm{n}},S_{1,\mathrm{n}},V_{1,\mathrm{n}},E_{1,\mathrm{n}},C_{1,\mathrm{n}}\\ D_{2,\mathrm{n}},{\mathrm{Q}}_{2,\mathrm{n}},S_{2,\mathrm{n}},V_{2,\mathrm{n}},E_{2,\mathrm{n}},C_{2,\mathrm{n}}\\ D_{3,\mathrm{n}},{\mathrm{Q}}_{3,\mathrm{n}},S_{3,\mathrm{n}},V_{3,\mathrm{n}},E_{3,\mathrm{n}},C_{3,\mathrm{n}}\end{array}\right]$$

(6)

Thirdly, the weights of each indicator were calculated. For each combination of traffic flow, the weights of the six indicators are different:

$$X_i=\left[D_{\mathrm{i},\mathrm{n}},{\mathrm{Q}}_{\mathrm{i},\mathrm{n}},S_{\mathrm{i},\mathrm{n}},V_{\mathrm{i},\mathrm{n}},E_{\mathrm{i},\mathrm{n}},C_{\mathrm{i},\mathrm{n}}\right]$$

(7)

(1)

The matrix M was expanded, where M represents the A_n matrix, and each element in A_n is represented by $m_{jk}$, where j represents the j-th scheme and k represents the k-th evaluation indicator.

$$\mathrm{M}=\left[{\mathrm{m}}_1{,\mathrm{m}}_2,\dots ,{\mathrm{m}}_k,\dots ,{\mathrm{m}}_6\right]$$

(8)

In matrix M, each vector m_k:

$${\mathrm{m}}_k=\left[{\mathrm{m}}_{1k}{,\mathrm{m}}_{2k},{\mathrm{m}}_{3k}\right]$$

(9)

The six evaluation indicators (delay, queue length, parking frequency, vehicle quantity, conflicts frequency, and CO emissions) are represented by k = 1–6. Among the first five indicators, the higher the indicator value, the better the solution effect. Therefore, in order to unify the evaluation method, it is necessary to normalize the six indicators, as shown in the equation.

$${{\mathrm{m}}^{\prime }}_k=\left[\begin{array}{l}\max \left\{{\mathrm{m}}_{1k}{,\mathrm{m}}_{2k},{\mathrm{m}}_{3k}\right\}-m_{1k}\\ \max \left\{{\mathrm{m}}_{1k}{,\mathrm{m}}_{2k},{\mathrm{m}}_{3k}\right\}-m_{2k}\\ \max \left\{{\mathrm{m}}_{1k}{,\mathrm{m}}_{2k},{\mathrm{m}}_{3k}\right\}-m_{3k}\end{array}\right],k\ne 3$$

(10)

The processed matrix M′ can yield the following:

$$M^{\prime }=\left[{{\mathrm{m}}^{\prime }}_1{,{\mathrm{m}}^{\prime }}_2,{\mathrm{m}}^{\prime }{}_3,\dots ,{\mathrm{m}}^{\prime }{}_6\right]$$

(11)

(2)

In addition to normalization, it is also necessary to standardize the different measurement units of these six indices. Finally, the comprehensive indicator (i.e., converting absolute indicator to relative indicator) is obtained as the weight indicator. The standardization process formula is as follows:

$$m^{″}{}_{jk}=\frac{m^{\prime }{}_{jk}-\min \{m^{\prime }{}_{1k},\dots ,m^{\prime }{}_{3k}\}}{\max \{m^{\prime }{}_{1k},\dots ,m^{\prime }{}_{3k}\}-\min \{m^{\prime }{}_{1k},\dots ,m^{\prime }{}_{3k}\}}$$

(12)

(3)

The weight of indicator-k at the n-th traffic volume situation is as follows:

$$p_{jk}=\frac{m{{}^{″}}_{jk}}{{\displaystyle \sum _{j=1}^{e}m{{}^{″}}_{jk}}},j=1 \mathrm{to} 4,k=1 \mathrm{to} 6$$

(13)

The indicator-k’ entropy value:

$$e_k=-\lambda {\displaystyle \sum _{j=1}^e{p}_{jk}\ln (p_{jk})}$$

(14)

The computation for redundancy in information entropy is outlined as follows:

$$d_k{=1-\mathrm{e}}_k$$

(15)

Equation (15) was used for the weight of each indicator 15 calculations.

$$w_k=\frac{d_k}{{\displaystyle \sum _{k=1}^f{d}_k}}$$

(16)

The weights of six evaluation indicators were recorded as the following:

$$w_n=\left[w_{n,1},w_{n,2},\dots ,w_{n,\mathrm{k}},\dots ,w_{n,6}\right]$$

(17)

The weights of the nth traffic and the last 49 traffic combinations in matrix form were combined as the following:

$$\mathrm{W}=\left[{\mathrm{W}}_1,{\mathrm{W}}_2,\dots ,{\mathrm{W}}_n,\dots ,{\mathrm{W}}_{49}\right]$$

(18)

6.3. Evaluation of Schemes

The design schemes were evaluated comprehensively under 49 traffic volume situations.

(1)

The ratio of the scheme-j value to all three scheme values:

$$p^{}{}_{jk}=\frac{m{{}^{″}}_{jk}}{{\displaystyle \sum _{j=1}^{e}m{{}^{″}}_{jk}}},j=1 \mathrm{to} 4,k=1 \mathrm{to} 6$$

(19)

(2)

The score of indicator-k in scheme-j was recorded as the following:

$$z_{jk}{=\mathrm{p}}_{jk}\times {\mathrm{W}}_n{}^T,k=1 \mathrm{to} 6$$

(20)

(3)

The score of the calculation formula for the scheme-j is the following:

$$z_j={\displaystyle \sum _{k=1}^6{z}_{jk}}$$

(21)

(4)

The three schemes’ final scores are shown in Z_n as follows:

$$Z_n={\left[z_1,z_2,z_3\right]}^T$$

(22)

Finally, the scores of all transportation combinations in matrix Z were summarized.

After obtaining matrix Z, the highest score was located to find the best solution for each traffic combination condition, and Matrix T was obtained based on the 7 × 7 traffic combination. The optimal solution result under 7 × 7 traffic volume conditions, namely matrix T, is shown in the following figure:

From Figure 6, it can be seen that Scheme 1 is distributed in the bottom left corner of the matrix, which means that the original construction area scheme can be evaluated with higher scores when the traffic volume on the two directions is low. With the gradual increase in traffic volume at the eastbound and the westbound, Scheme 2 shows better performance. However, as the traffic volume in the east direction further increases, the advantages of Scheme 3 in indicators such as traffic delay and traffic conflict become more apparent when it exceeds 1422 (veh/h). Moreover, due to the wider lane layout on the eastbound of Scheme 3, Scheme 2 has no advantage when the eastbound traffic volume exceeds 1422 (veh/h). The optimal performance ratio for Schemes 1, 2, and 3 is 6:19:24.

Based on the results of the EWM, Scheme 1 is optimal when the traffic volume on both entrances does not exceed 1000 (veh/h). But, when the westbound traffic volume exceeds 1400 (veh/h), Scheme 2 shows better traffic performance. However, Scheme 2 has a relatively low level of transportation service for the east entrance. As the eastbound traffic volume exceeds 1422 (veh/h), Scheme 3 exhibits an absolute advantage in all transportation combinations. The optimal performance ratio for Schemes 1, 2, and 3 is 6:19:24.

7. Conclusions

With the acceleration of China’s urbanization process, the transportation network of cities is becoming increasingly dense. The original road network in the city can no longer bear the current traffic pressure, and the main measures to deal with urban congestion are to renovate and expand existing roads or build subways and suspension rails for transportation. Therefore, it is inevitable that there will be more construction work zones on urban roads, and the establishment of construction work zones at signalized intersections makes intersections, which are already bottlenecks in the urban road network, more congested, causing adverse effects on citizens and regional economic development. For a rapidly developing city, if long-term traffic congestion is created to alleviate future traffic pressure, it will not only be difficult to meet the travel needs of citizens but also violate the current concept of sustainable development worldwide, increasing motor vehicle emissions.

Therefore, in order to alleviate the traffic pressure caused by the construction work zone on the intersection, the different forms of lane occupation in the suspension railway work zone were analyzed for different traffic volumes. This study established three schemes for different occupation forms and positions of work zones based on measured data and simulated them using VISSIM (2022 student). Then, this paper adopts a unique multiobjective decision-making method to evaluate the impact of different work zone occupancy positions on the operation of signalized intersections. In order to ensure the accuracy and comprehensiveness of the simulation results, five typical operational efficiency indicators and one safety indicator were selected for simulation modeling. Then, an evaluation table EWM was conducted for different suspension railway workspace locations. Based on the results of the data analysis, the following conclusions can be drawn.

Each of the three schemes has its own optimal performance under different traffic conditions. Scheme 1 (original scheme) can be used for situations with low traffic volume at the west or east entrance. Scheme 2 is mainly suitable for situations with high traffic volume in the opposite lane, and Scheme 3 is mainly suitable for situations with high traffic volume. However, when the traffic volume at the east entrance is greater than 2133 veh/h, no solutions can solve the traffic problem, and all six indicators have seriously deteriorated. If different schemes are selected for the construction of a full-range suspension railway based on different traffic conditions, it is of great significance to reduce the driving interference, parking frequency, and exhaust emissions of vehicles at the intersection in the work zone.

The entropy weight evaluation method model established in this article can quantify the impact of different suspension railway work zone locations on traffic operations under different traffic volumes. Furthermore, it reduces the degree of reduction in intersection traffic efficiency caused by the layout of work zones. As for Schemes 2 and 3, when the traffic volume in a certain direction of the intersection exceeds a certain threshold, the work zone placed in the opposite lane or in the center of the road should be considered to achieve better traffic efficiency. At present, various cities in China are constructing monorails, which is bound to generate a large number of construction work zones. For countries with relatively mature monorail development, such as Germany, France, the United Kingdom, and Japan, this article proposes a work zone evaluation method for the impact of work zones on intersections, which can also be applied to the expansion and maintenance projects of monorails. This evaluation method is based on the entropy weight method, which avoids the problem of subjective judgment by humans while considering traffic conflicts and traffic efficiency and regards carbon dioxide emissions as an important evaluation indicator for urban carbon neutrality. The research results can be used to guide transportation personnel to determine different road planning schemes, which is also an attempt to reduce the impact of work zones on traffic, reduce traffic congestion, and exhaust emissions through technical means.