Jurnal Sistem dan Manajemen Industri Vol 4 No 2 December 2020, 129-136

Available online at: http://e-jurnal.lppmunsera.org/index.php/JSMI Jurnal Sistem dan Manajemen Industri

ISSN (Print) 2580-2887 ISSN (Online) 2580-2895

Implementation of connection scan algorithm in intermodal transportation : a case study

Vertic Eridani Budi Darmawan1*, Yuh Wen Chen2 1 Department of Industrial Engineering, Universitas Negeri Malang, Jalan Semarang No. 5, Malang 65145, Indonesia 2 Institute of Industrial Engineering and Management, Da Yeh University, No. 168 University Rd. Dacun, Changhua 51591, Taiwan

ARTICLE INFORMATION ABSTRACT

Article history: Accessibility to tourist destinations is an important component in a

Received: November 26, 2020 tourism system, especially for natural tourist destinations located in Revised: December 24, 2020 suburban areas. Good linkage of information and physical Accepted: December 28, 2020 connections with local transportation services for intercity travel can facilitate more people to travel and promote national tourism Keywords: destinations. This research takes the popular national tourism destinations and their public transportation service in Taiwan as a Journey planner research object due to the unavailability of integrated public Intermodal transportation information service. Free Independent Travelers (FIT) Tourism demand is growing. This research aims to integrate intermodal public Connection scan algorithm Free independent travelers transportation information to support FIT by proposing a seamless way journey planner. In this scenario, the journey planner requires timetable data as input. The Connection Scan Algorithm is used to find the earliest arrival time routes at their destinations. This journey planner is built in PHP language and can complement the official tourism travel information website by Tourism Bureau, MOTC. Hence, the FIT could get the quickest routes to reach the destinations without compiling the public transportation information provided independently.

This is an open-access article under the CC–BY-NC-SA license. *Corresponding Author

Vertic Eridani Budi Darmawan E-mail: [email protected]

© 2020 Some rights reserved

1. INTRODUCTION the deterioration of the natural, social, and cultural The rapid growth of tourism lies in the high environment [1]–[3]. Tourism exacts a conse- trend towards globalization, and the world quence of change, and it’s required to be well continues to be more connected. The positive planned and managed. The successful tourism effects of globalization on tourism supply and development depends on maintaining a delicate demand are increasing economic prosperity, balance, which is to get a full benefit to outweigh creating employment opportunities, spreading the the cost and lessen the detrimental impacts as a usage of Information Communication and part of the change. It should be done in a Technology (ICT), developing new tourism sustainable way of economic development [4]. markets, and changing social structure [1]. Tourism could not possibly occur without Meanwhile, it also leads to an adverse impact on transportation, and it’s the most profound in the

http://dx.doi.org/10.30656/jsmi.v4i2.2772 129

Jurnal Sistem dan Manajemen Industri Vol 4 No 2 December 2020, 129-136

growth of global interconnectedness [5]. As a key travel (FIT). The accessibility of transportation element of the tourist experience, transportation between origin and destination will shape how provides the main connection between origin and tourist behaviors, especially whether using public destination tourism area and relieves tourists' or private transportation [6] and the cooperation of mobility [5], [6]. The accessibility of transpor- transportation modes, lead to improved transpor- tation between origin and destination, will shape tation experience and efficiency of the transpor- how tourist behaviors, especially whether they use tation systems [12]. Taiwan has a high-quality public or private transportation [6]. Concurrently, public transportation system [13]. The develop- providing the information to motivate tourists to ment of an intermodal passenger transportation take public transportation at the tourism destina- system continues to be improved [14]. It provides tion area, is one of 21 Tourism Agenda in Sustain- an excellent context for integrating an existing able Tourism [7]. public transportation system to the tourism desti- In Taiwan's case, most international tourists nation by developing the journey planner. would rather be Free Independent Travelers (FIT) The behavior of persons using public [8]. Most of them also searched about Taiwan’s transportation is different from traveling in private tourism information through the before mode. The persons are only able to depart and they arrived in Taiwan. After they came, the trans- arrive at the vehicle at specific stations within the portation information (60%) is the most informa- networks. Consequently, they need an alternative tion they were searching for, then followed by to find routes from their origin to the destinations. scenic spots (49%) and information The availability of routes in the public trans- (49%) [8]. The awareness of travel information- portation network depends on the timetable data. seeking behaviors could be as a prospecting guide It raises the timetable information problems by for the government agencies and businesses to finding the earliest arrival routes at their desti- decide an effective transportation service delivery nations within the network [15]. for the tourists [9]. Over the years, many research types contri- Given the increasing number of tourists, buted to algorithms to solve this problem some initiative has been made to publicize public [16]. Two major approaches have been proposed transportation used by tourists in Taiwan, to solve the timetable-based shortest path especially in the urban area, such as Fun Travel in problem: graph-based and non-graph approaches. Taipei (FTiT). Traveling across Taiwan (intercity The graph-based process is extended Dijkstra’s travel) could be convenient too because Taiwan’s algorithms with the two common methods called Government also offers innovative and intimate the time-dependent and time-expanded [15]. This travel service apps called “Tour Taiwan” and approach could not give convincing results or lead “VZTaiwan”. Both apps cover tourism desti- to high query times when applied to public nations all over Taiwan. They can also suggest a transportation networks with a different structure planned trip depends on tourist preference as the than the road network [17]. day's limit, the theme of the , and itinerary The first non-graph-based approach was pace. RAPTOR (Round-based Public Transit Routing) However, these services still lack reliable [18]. RAPTOR utilized the essential public transit connection information to the public mode traffic, network timetable elements, and it has no need and the service apps build independently with the preprocessing process and works in fully dynamic main tourism website. Meanwhile, the advanced scenarios. Currently, RAPTOR’s algorithm is traveler information service is needed to help FIT behind software such as Open Trip Planner [16]. decide to reach the tourism destination with No longer after RAPTOR algorithm was absolute certainty of transportation information invented, the Connection Scan Algorithm (CSA) [10]. The intercity traveling in Taiwan has was introduced in 2013 [19]. CSA uses an depended much more on public transportation approach of modeling the timetable data as a service [11]. directed acyclic graph or elementary connection. Providing tourism information services, es- Topologically, the primal connection is sorted by pecially transportation information, is critical in departure time, and the shortest path algorithm the globalization of the tourism industry. The only needs to scan all the connections of the tourism service shapes new tourism markets that timetable in a single array once. Currently, CSA is tourists become more independent and freer to behind the mobiTopp, a modular agent-based

130 http://dx.doi.org/10.30656/jsmi.v4i2.2775

Jurnal Sistem dan Manajemen Industri Vol 4 No 2 December 2020, 129-136

Travel Demand [20] and used in open data and halt to leave and enter the vehicle. initiatives for the public transportation route 푆 = (푠 , 푠 , 푠 , …, 푠 ) (2) planning ecosystem [21]. 1 2 3 푛 The algorithms are often built to answers a Time, when the vehicle arrives and departs at slightly different question and are specially a station, is well known. The vehicle arrives at implemented on top of other data models. Instead each station (s) at arrival time 푡푎푟푟 and departs at of developing a new algorithm, this research is the departure time 푡푑푒푝. The times are fully looking for trouble-free implementation with the described in the timetable. smallest building block needed to fit our needs. The foundational element in the timetable is CSA is slightly more efficient than RAPTOR, but a connection (푐). It represents a vehicle that drives it is substantially straightforward [16]. Hence, in between one stop to another stop without any this research, CSA was selected as the proven intermediate halt [22]. fastest algorithm to solve our problem in public Definition 2.2. transportation route planning than traditional route A connection is a tuple planning solutions based on variants Dijkstra’s algorithm [16], [20], [22]. 푐 = (푠푑푒푝, 푠푎푟푟, 푡푑푒푝, 푡푎푟푟, 푡푟) (3) The purpose of this study is to integrate where: intermodal public transportation information to 푠 is the departure station support FIT by proposing a seamless way journey 푑푒푝 푠 is the arrival station planner that has not been addressed yet. In this 푎푟푟 푡 is the departure time scenario, the journey planner requires timetable 푑푒푝 푡 is the arrival time data as input. The Connection Scan Algorithm is 푎푟푟 is a trip id used to find the earliest arrival time routes at their 푡푟 destinations. with 푠푑푒푝 ≠ 푠푎푟푟 and 푡푑푒푝 ≤ 푡푎푟푟.

By definition 2.2, the same trip id (푡푟) forms 2. RESEARCH METHODS a set of connections representing a vehicle's 2.1 Connection Scan Algorithm movement from the first departure station to the The Connection Scan Algorithm (CSA) was end arrival station. A sequence of stations includes firstly introduced in 2013 [19] and well-proven as the first and last stations along with corresponding an efficient public transportation route planning arrival time (푡푎푟푟) and departure time (푡푑푒푝) in each [22]. The elementary connection is a fundamental station defined as a trip. A trip that covers n unit to compute the earliest arrival time [22]. It is stations has (n-1) connections. In public the basic building block of the timetable, and CSA transportation, the vehicles that go on a trip efficiently arranges it into a single array. CSA (onward trip) and return to the starting station computes the network by scans the connection (return trip) are considered two different trips. array linearly to get the journey plan. The basic Definition 2.3. problem idea that CSA offer is the earliest arrival A trip is a sequence of connections time and extends to handle multi-criteria profile queries such as the earliest arrival time and the 푡푟푖푝 = (푐1, 푐2, 푐3, ... 푐푛−1) (4) minimum number of transfers [22]. where In CSA, the timetable for public 푐푖 = (푠푖, 푠푖+1, 푡푑푒푝(푖), 푡푎푟푟(푖+1), 푡푟) 푠 푡 transportation could define in a mathematical 푐푠푎푟푟 = 푐 푑푒푝 and 푐푡푎푟푟 < 푐 푑푒푝 for every i (5) formulation as a tuple: 푖 푖+1 푖 푖+1 The footpath is considered transfer time, (푆, 퐶, 푇, 퐹) (1) which is the time needed to walk between a nearby where station from one trip to another trip (trip transfer). S is a set of Stations Each footpath consists of two stations with an C is a set of Connections associate the duration of walking. In a simple CSA T is a set of Trips model, transfer time has been assumed to be zero. F is a set of Footpaths. Meanwhile, the fact is the transfer time is non- Definition 2.1. zero. The main characteristic of transfer time is A station is a place where the passenger can stand often dependent on a connection, the central hub

http://dx.doi.org/10.30656/jsmi.v4i2.2772 131

Jurnal Sistem dan Manajemen Industri Vol 4 No 2 December 2020, 129-136

or station where the transfer between vehicles occurs, and the needed time. The CSA algorithm is described in Fig. 1.

1: Input∶ Connection Stream (퐶푆), Station Stream (SS), query (푠dep, 푠arr, 푡푠푡푎푟푡) 2: Output∶ Itinerary (I) 3: Initialize: IC, d 4: for all 푠푖 ∈ SS 퐝퐨

5: IC [푠푖] ← 푛푢푙푙 6: 푑[푠푖] ←∞ 7: 푑(푠dep) ← 푡푠푡푎푟푡 8: end for 9: Main Loop 10: for all 푐 ∈ 퐶 퐝퐨 11: 퐢퐟 푡푑푒푝[퐶푖] ≥ 푑[푠푑푒푝[퐶푖]] 퐚퐧퐝 12: 푡푎푟푟[퐶푖] < 푑 [푠푎푟푟[퐶푖]] 풕풉풆풏 13: 푑 [푠푎푟푟[퐶푖]] ← 푡푎푟푟[퐶푖] 14: 퐼퐶 [푠푎푟푟[퐶푖]] ← 퐶푖 Fig. 2. Tourism journey planner workflow 15: end if These data are as an input to the Algorithm 16: end for Implementation Simulation. In this stage, the 17: 퐼 ← 퐼푡푖푛푒푟푎푟푦 [퐼퐶] timetable needs to formalize and encode what 18: return 퐼 transportation mode exists, the station where they depart and arrive when they depart and arrive, and Fig. 1. Algorithm CSA [22] how the traveler transfer between transportation 2.2. Methodology modes. The timetable formalization process The initial stage of this research is defining consists of extracting and transforming data. Once the problem. The problem statement is the un- the algorithm simulation worked, the next step is availability of integrated public transportation the journal planner design and implementation. information to support FITs to reach tourism This stage consists of developing and testing the destination. This problem leads to the objective system’s software. As we addressed this system to and scope of this research: as building a seamless be able to complement existing tourism website, way journey planner that could attach to the the system implemented in “PHP”, “HTML”, existing official website with the scopes are focus “CSS”, “JavaScript” as a programming language on international tourists who enter Taiwan from and MySQL as database service to store the Taiwan International (this considered as a transportation timetable data. The research work- gateway or main hub) and the top three tourism flow of tourism intermodal transportation journey destination, namely Jiufen (九份), Sun Moon planner is shown in Fig. 2. Lake (日月潭), and Kenting National Park (墾丁 國家公園). The third stage is collecting the data. The data needed in this research is the timetable data, which have a trip to the selected tourism destination. The public transport timetable data involve three transportation modes: Taiwan High-Speed Rail (THSR), Taiwan Railways Administration, and Intercity . Like nature, these lead to data integration among three different transportation modes called intermodal Fig. 3. Tourism intermodal transportation transportation. journey planner system architecture

132 http://dx.doi.org/10.30656/jsmi.v4i2.2775

Jurnal Sistem dan Manajemen Industri Vol 4 No 2 December 2020, 129-136

The tourism intermodal transportation b. Transforming Data journey planner architecture in Fig. 3 shows how In the second phase, the collected data the journey planner implementation process from transformed into a spreadsheet, and adding the collecting the data through the journey planner transfer path's connection with the total for real works. The users or tourists used the web connections is 1711. After completing the application as the user interface and inputted the spreadsheet data and importing it to the database origin station, destination station, and departure form, the database functionality was added to the time. In the backend, the CSA proceeds the input web page. and gives the output as the fastest journey (the 3.2 Algorithm Implementation earliest arrival time route) back to the web Before going through the system implemen- application. tation, the basic form algorithm implementation is

used as simulation. Start with the formalization of 3. RESULTS AND DISCUSSION the table and assumption that used. The table 3.1 Timetable Formalization format that we need is the connection stream and The journey planner works with a data struc- station stream. In a simple connection model, the ture previously built from each timetable transpor- transfer time has been assumed to be zero. The tation data to the CSA timetable. The timetable most straightforward problem that we consider to transportation data is being processed and solve is the earliest arrival problem given the input transformed on a desktop computer. The processes of timetable, origin station (푠표푟푖), destination consist of two phases, namely extracting data and station (푠푑푒푠), and starting time at origin station transforming data. (푡푠푡푎푟푡) to find the fastest way to arrive at the destination.

Fig. 5. Transportation network for initial result The transportation network describes in Fig. 5. There are four routes and eight trips in the transportation network with the different public transportation services represented by the net- work's different colors. The first route is Taoyuan - Taipei, the second route is Taipei - Ruifang, the third route is Ruifang - Jiufen, and the last route is Taipei – Puli – Sun Moon Lake. By looking at the Fig. 4. The intermodal transportation network examples network, Taipei Station could be catego- a. Extracting Data rized as a transit station. It handles a substantial In the first phase, data were collected and amount of traffic and connects the elements of the extracted manually from each transportation mode same of different transportation networks in scale. which already identified. While pulling the data, An example of the problem is the fastest way to the transportation network was built to determine Sun Moon Lake from Taoyuan, leaving at 06.00. the transit or transfer station. It helps to define the Building connections are the fundamental unit transfer path between transformation mode. The building block to calculate the earliest arrival time intermodal transportation network shows in Fig. 4. given that timetable. The amount of connection of

http://dx.doi.org/10.30656/jsmi.v4i2.2772 133

Jurnal Sistem dan Manajemen Industri Vol 4 No 2 December 2020, 129-136

each trip could be calculated by subtracting 1 from planning, the user can set parameters that are the amount of station, e.g., trip 07 and trip 08 have considered by the route planner. The parameters 3 stations, then the number of connections is 3 – 1 are the departure station, arrival station, and = 2. For trip 07, the connection would be like leaving time. For example, the user will go from below: Airport Terminal 2 Station (A13) as departure station to Kenting (Xiaowan) as arrival station and 푐1 = (푡푎푖푝푒푖 푠푡푎푡푖표푛 , 푝푢푙푖 푠푡푎푡푖표푛 , 0630 , 푑푒푝 푎푟푟 푑푒푝 at 9.14 am as leaving time (Fig. 8). 0830푎푟푟, 푡푟07)

푐2 = (푝푢푙푖 푠푡푎푡푖표푛푑푒푝, 푠푢푛푚표표푛 푙푎푘푒 푠푡푎푡푖표푛푎푟푟, 0830푑푒푝, 0900푎푟푟, 푡푟07) The connection (C) stream sorted by departure time shows in Fig. 6. The basic idea how the CSA works is by keeping the best arrival time 푑 (earliest arrival time) and arrival connection 퐼퐶 (InConnection) for every station (s) to get the itinerary leaving from origin station (푠표푟푖) to destination station (푠 ) at starting time (푡 ). 푑푒푠 푠푡푎푟푡 Fig. 8. Journey planner form The journey planner lists the complete journey planner for each station, time, and transfer path between transportation modes. The transfer path helps the user to understand how many transportation modes they will use for the chosen journey. The transfer path is calculated using the best knowledge of the footpath. It shows in Fig. 9.

Fig. 6. Sorted connection stream The connection scan iterates all the connec- tions to see whether it can improve any connec- tion. The earliest arrival time from the origin station to any other station is computed by exploring all the connections. The backtracking search is utilized from the destination station and Fig. 9. The tourism intermodal transportation rally on the IC (InConnection) until it reaches the journey planner result origin station. This is shown in Fig. 7. 4. CONCLUSION This research is providing tourism intermodal transportation journey planner in a study case of Taiwan tourism. The journey planner is success- fully built in PHP language embedded in the existing Taiwan Tourism Website. Our efforts will Fig. 7. The itinerary initial result help the international tourists get full information After the simulation, the system implemen- about the trip journey as an idea of a seamless tation process is integrated with the database and connection. web application as the user interface. A journey This research does not limit to selected planner's main function is to plan one route from a transportation data. It is the beginning in demon- starting point to a destination. Before the route strated the whole process of implementing the

134 http://dx.doi.org/10.30656/jsmi.v4i2.2775

Jurnal Sistem dan Manajemen Industri Vol 4 No 2 December 2020, 129-136

journey planner. The scalability, manpower, and 10.1016/j.jtrangeo.2006.12.008. government or service companies can gather [7] UNEP, Tourism and Local Agenda 21 The precise and updated data in the future. For Role of Local Authorities in Sustainable example, the data could be dynamic due to some Tourism. 2003. Available: tradeoffs, such as the unique transportation table https://wedocs.unep.org/handle/20.500.11 due to some special events (Chinese New Year, 822/7920. Summer Holiday, or Winter Holiday). [8] T. T. Bureau, “2007 Annual Survey Report We expect to add some features to increase on Visitors’ Expenditure and Trends in user value in practice later. For example, we may Taiwan.” Ministry of Transportation and cooperate with transportation service operators or Communication Taiwan, 2008. Available: government transportation institutions to increase https://admin.taiwan.net.tw/Handlers/File the data variety. We may add a function of mini- Handler.ashx?fid=0d95ed6e-93f5-4ffa- mum transfer pattern [22], add a function of afff-c61c9e080982&type=4&no=1. transfer pattern combined footpath, cab, and bike [9] Z. Kambele, G. Li, and Z. Zhou, as a transportation method in the country [23]. We “Travelers’ Information-Seeking may add a multiple-day trip [24] and add some art Behaviors,” J. Travel Tour. Mark., vol. 32, features to get a fascinating user interface design. no. 1–2, pp. 141–152, Feb. 2015, doi: 10.1080/10548408.2014.986017. REFERENCES [10] C.-M. Feng, “New prospects of [1] L. Dwyer, “Globalization of Tourism: transportation mobility,” IATSS Res., vol. Drivers and Outcomes,” Tour. Recreat. 38, no. 1, pp. 22–26, 2014, doi: Res., vol. 40, no. 3, pp. 326–339, 2015, doi: 10.1016/j.iatssr.2014.05.005. 10.1080/02508281.2015.1075723. [11] L. W. Lan, M.-T. Wang, and A. Y. Kuo, [2] J. W. Lee and T. Brahmasrene, “Development and Deployment of Public “Investigating The Influence of Tourism Transport Policy and Planning in on Economic Growth and Carbon Taiwan,” Transportation (Amst)., vol. 33, Emissions: Evidence from Panel Analysis no. 2, pp. 153–170, 2006, doi: of The European Union,” Tour. Manag., 10.1007/s11116-005-3048-0. vol. 38, pp. 69–76, 2013, doi: [12] M. Efthymiou and A. Papatheodorou, 10.1016/j.tourman.2013.02.016. “Intermodal passenger transport and [3] D. Buhalis and R. Law, “Progress in destination competitiveness in Greece,” Information Technology and Tourism Anatolia, vol. 26, no. 3, pp. 459–471, Jul. Management: 20 years on and 10 years 2015, doi: after The Internet-The State of eTourism 10.1080/13032917.2015.1012171. Research,” Tour. Manag., vol. 29, no. 4, [13] Y. Shi and X. Yang, “The Public pp. 609–623, 2008, doi: Transportation System of High Quality in 10.1016/j.tourman.2008.01.005. Taiwan,” Procedia - Soc. Behav. Sci., vol. [4] S. Page, Transport and Tourism: Global 96, pp. 1350–1361, 2013, doi: Perspectives. Pearson Prentice Hall, 2009. 10.1016/j.sbspro.2013.08.153. Available: [14] P. Chang and H.-S. Shieh, “Seamless https://books.google.co.id/books?id=yKlX Service Strategies for Passenger PgAACAAJ. Transportation in Taiwan,” Int. J. Bus. Soc. [5] G. Moscardo and P. L. Pearce, “Life Cycle, Sci., vol. 7, no. 6, pp. 136–148, 2016. Tourist Motivation and Transport: Some Available: Consequences for the Tourist Experience,” https://ijbssnet.com/journals/Vol_7_No_6 in Tourism and Transport, First Edit., _June_2016/13.pdf. Elsevier, 2004, pp. 29–43, doi: [15] M. Müller-Hannemann, F. Schulz, D. 10.1016/B978-0-08-044172-6.50007-8. Wagner, and C. Zaroliagis, “Timetable [6] W. Gronau and A. Kagermeier, “Key Information: Models and Algorithms,” in Factors for Successful Leisure and Algorithmic Methods for Railway Tourism Public Transport Provision,” J. Optimization, Berlin, Heidelberg: Springer Transp. Geogr., vol. 15, no. 2, pp. 127– Berlin Heidelberg, 2007, pp. 67–90, doi: 135, 2007, doi: 10.1007/978-3-540-74247-0_3.

http://dx.doi.org/10.30656/jsmi.v4i2.2772 135

Jurnal Sistem dan Manajemen Industri Vol 4 No 2 December 2020, 129-136

[16] H. Bast et al., “Route Planning in Verborgh, “Decentralised open data Transportation Networks,” in Algorithm publishing for the public transport route engineering, Springer, 2016, pp. 19–80, planning ecosystem,” Situating Open Data doi: 10.1007/978-3-319-49487-6_2. Glob. Trends Local Context, pp. 155–159, [17] H. Bast, “ or Public Transport—Two 2020. Available: Worlds,” in Efficient Algorithms, Springer, https://biblio.ugent.be/publication/867824 2009, pp. 355–367, doi: 10.1007/978-3- 7/file/8678249. 642-03456-5_24. [22] J. Dibbelt, T. Pajor, B. Strasser, and D. [18] D. Delling, T. Pajor, and R. F. Werneck, Wagner, “Connection Scan Algorithm,” “Round-Based Public Transit Routing,” ACM J. Exp. Algorithmics, vol. 23, pp. 1– Transp. Sci., vol. 49, no. 3, pp. 591–604, 56, Nov. 2018, doi: 10.1145/3274661. Aug. 2015, doi: 10.1287/trsc.2014.0534. [23] T. Prasad, K. Sathyanarayanan, S. Tiwari, [19] J. Dibbelt, T. Pajor, B. Strasser, and D. N. Goveas, and B. Deshpande, “t-CSA: A Wagner, “Intriguingly Simple and Fast fast and flexible CSA implementation,” in Transit Routing,” in International 2016 8th International Conference on Symposium on Experimental Algorithms, Communication Systems and Networks Springer, 2013, pp. 43–54. doi: (COMSNETS), Jan. 2016, pp. 1–6, doi: 10.1007/978-3-642-38527-8_6. 10.1109/COMSNETS.2016.7439937. [20] L. Briem et al., “Integrating public [24] W. Wisittipanich and C. Boonya, “Multi- transport into mobiTopp,” Futur. Gener. objective Tourist Trip Design Problem in Comput. Syst., vol. 107, pp. 1089–1096, Chiang Mai City,” IOP Conf. Ser. Mater. 2020, doi: 10.1016/j.future.2017.12.051. Sci. Eng., vol. 895, no. 1, pp. 1–7, Jul. [21] J. Rojas, B. Marcelis, E. Vlassenroot, M. 2020, doi: 10.1088/1757- van Compernolle, P. Colpaert, and R. 899X/895/1/012014.

136 http://dx.doi.org/10.30656/jsmi.v4i2.2775