Developing Hybrid Simulation Model to Improve Road Traffic Management

Al Aziz, O. R., et al., International Journal of Innovation in Enterprise System Vol. 04 No. 01 (2020) p. 56-65 *Corresponding author: hayati.asih@ie.uad.ac.id | 57 Fig. 1 – Average delay in intersection in Yogyakarta[1] Pelemgurih intersection has four sections of road, there are Patukan road (red line) on the west, Yogyakarta-Wates road (blue line) on the east, North Siliwangi road (orange line) on the north, and South Siliwangi road (green line) on the south. It is as shown in Fig. 2. In defining the direction of each road, there are the dash line which means the vehicles must follow the traffic light, and the straight line which means the vehicle must go ahead, not follow the traffic light. For instance, it can be seen the direction from Yogyakarta-Wates road to South Siliwangi road must turn left directly, even the traffic light is red. U North (North Siliwangi Road) East (Yogyakarta-Wates Road) South (South Siliwangi Road) West (Patukan Road) Red : Patukan Road vechicle flow Go ahead Follow the traffic light Blue Orange Green : Yogya-Wates Road vechicle flow : North Siliwangi Road vechicle flow : South Siliwangi Road vechicle flow Color Explanation : Fig. 2 – Pelemgurih intersection The preliminary study was conducted to understand the characteristics of the system by doing initial observation. The observation is divided by three periods, i.e. morning (06.00 am – 08.00 am), noon (11.00 am – 13.00 pm), and afternoon (15.30 pm – 17.30 pm). The initial observation was conducted on weekdays (Monday to Friday) and weekend (Saturday and Sunday). In addition, this research concerns on motorcycle (MC), low vehicle (LV) which includes car, and high vehicle (HV) which includes truck and bus. The results of this initial observation was the number of vehicle on the weekdays in the morning and in the afternoon has significantly higher than on weekend (refer to Fig. 3). The main reason is many workers pass this intersection which makes the system more complicated. Hence, this research would focus on the weekdays in the morning and in the afternoon. Al Aziz, O. R., et al., International Journal of Innovation in Enterprise System Vol. 04 No. 01 (2020) p. 56-65 © Copyright by Directorate of Research and Community Service, Telkom University | 58 Fig. 2 – Total volume vehicle on weekdays and weekend The higher number of vehicles waiting in line in a traffic light will effect to higher emission and air pollution. In addition, a bad road traffic management will effect to higher number of accidents. Therefore, road traffic management must be planned efficiently and effectively in order to minimize the risk of accidents and environmental impacts [2]. This research developed hybrid simulation method with the objective is to minimize the number of vehicles waiting in lines. First of all, MKJI (“Manual Kapasitas Jalan Indonesia”) and Webster method were developed to determine the duration of green light accordance with the geometric, the condition of the intersection, capacity of the road, etc. Then, the simulation model was developed to evaluate the number of vehicles waiting in lines. Finally, the least amount of vehicles waiting in line was chosen. A case study will then be provided in Pelemgurih intersection located in Yogyakarta, Indonesia for demonstrating the applicability of the developed method. The results obtained can help the generating desired decision alternatives that will important for Department of Transportation, Indonesia to enhance the road traffic management with low number of vehicles waiting in line. The remainder of the present paper is organized as follows. Section 2 reviews the theoretical framework. Section 3 described the proposed methodology. Section 4 presents the results and discussion. Finally, section 5 concludes with a discussion of further research direction. 2. Theoretical Framework There are previous papers that were discussed about road traffic management. Kamran et al. [3] developed simulation modeling using Arena Simulation Software to determine the traffic light timing. There were several scenarios were proposed and the minimum average time spent in the system was chosen. Then, Tama et al. [2] also developed several scenarios based on various traffic light duration by using simulation model. The objective was to find the scenario that has shorter vehicle queue on the junction. In addition, Rahman [4] developed simulation model by using Vissim 8 program and MKJI method by changing the duration of traffic light, road widening, and improvement on the road geometric. The result showed the decreases on queue length, queue length maximum, and the number of vehicles stop. Next, Prihati [5] developed simulation model by using Arena simulation software to minimize the waiting time in account payment counter. Next, Yunata [6] also proposed Arena to simulate Simpang Empat Pontianak in order to number of vehicles waiting and waiting time could be reduced. In a comprehensive study of road traffic management, Harahap et al. [7] proposed a traffic simulation system, called by LINTAS using the SimEvents toolbox and runs on MATLAB-Simulink software to minimize the average queue length. Compare to another researches that develop simulation model for road traffic management, Fauzi et al. [8] developed Webster and Greenshield method to determine the green and red traffic light in Bandar Kidul intersection, Kediri, Indonesia. In a study conducted by Pornamasari et al. [9], it was shown that Webster method was proposed to calculate optimum cycle time in Babe Palar street intersection. This study set out to critically examine the ways in managing road traffic. The significant finding from previous researches is to calculate the green traffic light mostly using simulation model, MKJI and Webster method separately. Therefore, this research develops the hybrid simulation model by integrating simulation model, MKJI and Webster method to calculate the green traffic light more accurate. This study has objective to minimize the number of vehicles waiting in line. Al Aziz, O. R., et al., International Journal of Innovation in Enterprise System Vol. 04 No. 01 (2020) p. 56-65 *Corresponding author: hayati.asih@ie.uad.ac.id | 59 3. Methodology This research was conducted based on Pelemgurih intersection in Yogyakarta. It focuses on the weekdays in the morning and afternoon as this intersection as one of the main ways for people leave and go to work so that it has higher congestion. To minimize the congestion, the hybrid simulation was developed. First of all, the problem was identified and then the objective was proposed. Next, the data was collected such as inter arrival time, the velocity of vehicles, current traffic light duration, intersection geometry data, set up vehicles data, etc. Then, the experimental design 1 was calculated using MKJI and Webster method to provide the duration of green traffic lights. After that, the simulation model was developed to evaluate the number of vehicles waiting in line according to different duration of green lights from MKJI and Webster method as the experimental design 2. Finally, the duration of green traffic light with minimum number of vehicles waiting in line was chosen. It can be shown in Fig. 3. Problem Objective Data collection Experimental Design 1: Calculating the green lights using MKJI and Webster method Experimental Design 2: Developing the simulation model to evaluate the number of vehicles waiting in line Output The duration of green lights with minimum number of vehicles waiting in line Fig. 3 – Research design 3.1 MKJI and Webster method MKJI (Manual Kapasitas Jalan Indonesia) method was developed on 1997 to determine the duration of green traffic light [10]. Fig. 4 presents how the MKJI works in calculating the green duration for each intersection. On the other hand, Webster was developed on 1960s by F.V. Webster [9]. Fig. 5 presents step by step of Webster in calculating the the green duration for each intersection. F. V. Webster developed an equation to calculate the average delay in driving when approaching an intersection and also derived an equation to obtain the optimum cycle time that results in minimum vehicle delays. Vehicle delays Al Aziz, O. R., et al., International Journal of Innovation in Enterprise System Vol. 04 No. 01 (2020) p. 56-65 © Copyright by Directorate of Research and Community Service, Telkom University | 60 occur because the number of vehicles entering a junction is greater than the number of vehicles leaving the intersection. In calculating the duration of green traffic light, MKJI method considers intersection environmental conditions, but Webster method does not. Fill the Sig 1 Form: 1. Code (direction) 2. Environmental type (RES/COM) 3. Side barriers 4. Median 5. Slope 6. Turn left directly (Yes/No) 7. Park distance (m) 8. Width of road 9. Width of road (in) 10. Width of road for turn left directly 11. Width of road (out) Start Fill the Sig 2 Form: 1. Determine the amount of vehicle flow on each side of the intersection with each vehicle movement (straight/turn right/left) 2. Change the unit of the vehicle from unit/ hour to smp/hour by multiplying the number of vehicles by the value of all emp 3. Determine turn ratio 4. Determine the current of UM (non-vehicle) 5. Determine the UM / MV ratio Fill the Sig 3 Form: 1. Code (direction) 2. Determine the phase 3. Determine the type of approach (P / O) 4. Determine the turning ratio of the vehicle (pLTOR; pLT; pRT) of sig 2 5. Slope 6. Determine the number of vehicle which turn right (RT) for type O 7. Determine the effective width (We = Enter) 8. Determine the base current saturation (S0) 9. Determine city size (Fcs) 10. Side barriers (FSF) 11. Slope (Fg) 12. Parking (Fp) 13. Turning factor for right turn (Frt = 1.0 + Prt x 0.26) 14. Turning factor for left turn (Frt = 1.0 Plt x 0.16) 15. Adjustable saturation current value (S = So x Fcs x Fsf x Fg x Fp x Frt x Flt) 16. Input total of vehicle flow (Q) 17. Calculate the current ratio (FR = Q / S) 18. Determine IFR ( FR) 19 Calculate Phase Ratio (PR = FR / IFR) 20. cycle time before adjustment (Cua = (1.5 x LTI + 5) / (1 IFR) 21. Calculate the green light duration (gi = (Cua LTI) x PRi End Fig. 4 – The Procedure of MKJI in Calculating The Green Duration Al Aziz, O. R., et al., International Journal of Innovation in Enterprise System Vol. 04 No. 01 (2020) p. 56-65 *Corresponding author: hayati.asih@ie.uad.ac.id | 61 Determine the vehicle volume (in) Q = vehicle flow x the value of emp Start Determine the saturation flow (s = 525 x w) W = width of road Determine the level of traffic light (in) Y = Q/s Calculate lost time (L = (I a) + 1) I = Period between green A = yellow period (3 seconds) 1 = average time lost per phase (2 seconds) Determine the optimum cycle time (C0 = (1,5 +5 )/ (1 )) L = total lost time Y = Sum of y for all phases Calculate effective green time (g = (C0 L) / ) g = Effective green duration y = The level of traffic flow at each intersection C0 = cycle time L = Total lost time per cycle Y = Maximum number of y for all phases End Fig. 5 – The Procedure of MKJI in Calculating The Green Duration 3.2 Simulation model The simulation models are developed to evaluate the number of vehicles waiting in line with different the green light durations which are resulted from MKJI and Webster method. This research employs Arena simulation software to develop these simulation models. In developing the simulation model, some probabilistic data were considered. There are inter arrival vehicles (MC, LV, and HV), setup each vehicle before leave the line, the speed of each vehicles, etc. Fig. 6 presents the current simulation model. The verification and validation were firstly conducted before the scenarios were developed. Al Aziz, O. R., et al., International Journal of Innovation in Enterprise System Vol. 04 No. 01 (2020) p. 56-65 © Copyright by Directorate of Research and Community Service, Telkom University | 62 Patukan Road MC4 patuk1 to patuk lalin1 to jogja 4_1


Introduction
Yogyakarta is known as tourist city and educational city in Indonesia. There are many local and foreign tourist and also student who interested come to this city in each year. It results in high congestion in some intersections around Yogyakarta. High congestion has bad impact on time, money, and air pollution as well. According to annually evaluation from Department of Transportation of Yogyakarta on 2018 [1], there are average delay in some crowded intersections in Yogyakarta as presented in Fig. 1. It can be seen that Pelemgurih intersection has the highest average delay compare to others. It is caused by it connects the south ring road and the west ring road. It is also national road with environmental type is commercial which means many sales around. In addition, this intersection is one of the main ways for many workers pass this way which makes this intersection more crowded and road traffic management become challenging.

Fig. 1 -Average delay in intersection in Yogyakarta[1]
Pelemgurih intersection has four sections of road, there are Patukan road (red line) on the west, Yogyakarta-Wates road (blue line) on the east, North Siliwangi road (orange line) on the north, and South Siliwangi road (green line) on the south. It is as shown in Fig. 2. In defining the direction of each road, there are the dash line which means the vehicles must follow the traffic light, and the straight line which means the vehicle must go ahead, not follow the traffic light. For instance, it can be seen the direction from Yogyakarta-Wates road to South Siliwangi road must turn left directly, even the traffic light is red. Color Explanation :

Fig. 2 -Pelemgurih intersection
The preliminary study was conducted to understand the characteristics of the system by doing initial observation. The observation is divided by three periods, i.e. morning (06.00 am -08.00 am), noon (11.00 am -13.00 pm), and afternoon (15.30 pm -17.30 pm). The initial observation was conducted on weekdays (Monday to Friday) and weekend (Saturday and Sunday). In addition, this research concerns on motorcycle (MC), low vehicle (LV) which includes car, and high vehicle (HV) which includes truck and bus. The results of this initial observation was the number of vehicle on the weekdays in the morning and in the afternoon has significantly higher than on weekend (refer to Fig.  3). The main reason is many workers pass this intersection which makes the system more complicated. Hence, this research would focus on the weekdays in the morning and in the afternoon.

Fig. 2 -Total volume vehicle on weekdays and weekend
The higher number of vehicles waiting in line in a traffic light will effect to higher emission and air pollution. In addition, a bad road traffic management will effect to higher number of accidents. Therefore, road traffic management must be planned efficiently and effectively in order to minimize the risk of accidents and environmental impacts [2].
This research developed hybrid simulation method with the objective is to minimize the number of vehicles waiting in lines. First of all, MKJI ("Manual Kapasitas Jalan Indonesia") and Webster method were developed to determine the duration of green light accordance with the geometric, the condition of the intersection, capacity of the road, etc. Then, the simulation model was developed to evaluate the number of vehicles waiting in lines. Finally, the least amount of vehicles waiting in line was chosen. A case study will then be provided in Pelemgurih intersection located in Yogyakarta, Indonesia for demonstrating the applicability of the developed method. The results obtained can help the generating desired decision alternatives that will important for Department of Transportation, Indonesia to enhance the road traffic management with low number of vehicles waiting in line.
The remainder of the present paper is organized as follows. Section 2 reviews the theoretical framework. Section 3 described the proposed methodology. Section 4 presents the results and discussion. Finally, section 5 concludes with a discussion of further research direction.

Theoretical Framework
There are previous papers that were discussed about road traffic management. Kamran et al. [3] developed simulation modeling using Arena Simulation Software to determine the traffic light timing. There were several scenarios were proposed and the minimum average time spent in the system was chosen. Then, Tama et al. [2] also developed several scenarios based on various traffic light duration by using simulation model. The objective was to find the scenario that has shorter vehicle queue on the junction. In addition, Rahman [4] developed simulation model by using Vissim 8 program and MKJI method by changing the duration of traffic light, road widening, and improvement on the road geometric. The result showed the decreases on queue length, queue length maximum, and the number of vehicles stop. Next, Prihati [5] developed simulation model by using Arena simulation software to minimize the waiting time in account payment counter. Next, Yunata [6] also proposed Arena to simulate Simpang Empat Pontianak in order to number of vehicles waiting and waiting time could be reduced. In a comprehensive study of road traffic management, Harahap et al. [7] proposed a traffic simulation system, called by LINTAS using the SimEvents toolbox and runs on MATLAB-Simulink software to minimize the average queue length.
Compare to another researches that develop simulation model for road traffic management, Fauzi et al. [8] developed Webster and Greenshield method to determine the green and red traffic light in Bandar Kidul intersection, Kediri, Indonesia. In a study conducted by Pornamasari et al. [9], it was shown that Webster method was proposed to calculate optimum cycle time in Babe Palar street intersection.
This study set out to critically examine the ways in managing road traffic. The significant finding from previous researches is to calculate the green traffic light mostly using simulation model, MKJI and Webster method separately. Therefore, this research develops the hybrid simulation model by integrating simulation model, MKJI and Webster method to calculate the green traffic light more accurate. This study has objective to minimize the number of vehicles waiting in line.

Methodology
This research was conducted based on Pelemgurih intersection in Yogyakarta. It focuses on the weekdays in the morning and afternoon as this intersection as one of the main ways for people leave and go to work so that it has higher congestion. To minimize the congestion, the hybrid simulation was developed. First of all, the problem was identified and then the objective was proposed. Next, the data was collected such as inter arrival time, the velocity of vehicles, current traffic light duration, intersection geometry data, set up vehicles data, etc. Then, the experimental design 1 was calculated using MKJI and Webster method to provide the duration of green traffic lights. After that, the simulation model was developed to evaluate the number of vehicles waiting in line according to different duration of green lights from MKJI and Webster method as the experimental design 2. Finally, the duration of green traffic light with minimum number of vehicles waiting in line was chosen. It can be shown in Fig. 3.

MKJI and Webster method
MKJI (Manual Kapasitas Jalan Indonesia) method was developed on 1997 to determine the duration of green traffic light [10]. Fig. 4 presents how the MKJI works in calculating the green duration for each intersection. On the other hand, Webster was developed on 1960s by F.V. Webster [9].

Simulation model
The simulation models are developed to evaluate the number of vehicles waiting in line with different the green light durations which are resulted from MKJI and Webster method. This research employs Arena simulation software to develop these simulation models. In developing the simulation model, some probabilistic data were considered. There are inter arrival vehicles (MC, LV, and HV), setup each vehicle before leave the line, the speed of each vehicles, etc. Fig. 6 presents the current simulation model. The verification and validation were firstly conducted before the scenarios were developed.

Results and Discussion
The results of experimental design 1 was the different duration of green traffic lights according to MKJI and Webster method on the weekdays in the morning and in the afternoon as presented in Table 1 and Table 2, respectively. For instance, in Table 1, North Siliwangi roadway has the current duration of green light on the weekdays in the morning is 25 seconds. After calculated with MKJI method and Webster method, the durations of green light are 25 seconds and 10 seconds, respectively. Generally, this result shows the duration of green lights from Webster method provides lower than MKJI method.  Then, the experimental design 2 was conducted to evaluate the number of vehicles waiting in line for different duration green traffic lights that are outputs from MKJI and Webster method. As seen in Fig. 12, Webster method has lower number of vehicles waiting line in the morning and afternoon rather than MKJI method. This is due to the long duration of the green traffic light resulted by MKJI method so that the traffic light cycle becomes longer and it affects the number of vehicles waiting in line which is higher than Webster method.

Conclusion
This study developed hybrid simulation method to manage road traffic, especially in determining the green traffic light. The objective is to minimize the number of vehicles waiting in line. This method incorporates MKJI, Webster method, and simulation method. First of all, the MKJI and Webster method were developed to calculate the green traffic light duration. Then, the simulation model was developed to evaluate the number of vehicles waiting in line according to different duration of green traffic lights from MKJI and Webster method.
The result shows that the duration of green traffic lights calculated by Webster method provides lower number of vehicles waiting in line. It is due to the short duration of green traffic light resulted by Webster method so that the traffic light cycle becomes shorter and it affects the number of vehicles waiting in line which is lower than MKJI method.
In general, the proposed method has advantages for Department of Transportation, Indonesia in improving the road traffic management in order to low number of vehicles waiting in line. This research has opportunity to be developed by considering distance among vehicles and proposing the other methods.