Application Design in Determining the Shortest Route Using Tabu Search Algorithm

This research designed a website application using Tabu Search Algorithm. We applied the Queuing method to find the shortest distance. Applications with this algorithm can provide the fastest route in the experiment of searching the shortest distance. We conducted this research by applying the examination in Rote Island, Indonesia. On the island, we examined the shortest distance from a starting point to another tourism place. In another side, this system utilized Google Map services to retrieve data about the distance between locations to another location with driving mode. The Google Maps service application is received in the form of translated JavaScript Object Notation (JSON), in order to be used as data and parameters in executing tabu search algorithms. The system was built using web-based through Hypertext Preprocessor (PHP) programming language, HyperText Markup Language (HTML), and Javascript, so it can be accessed and run by all devices that run the browser.


Introduction
Problems that often arise in finding the shortest distance or the fastest route in any area is the lack of information itself.It has also happened in the tourism activity, when people, in this case, tourists, need to find a path to get tourist attractions place or to know the level of a street crowded and the volume of the path.Due to their limited time, the shortest route search is important (Fard & Akbari, 2013).
Searching the shortest route is to determine the most optimal path, that is the path with the shortest route and the smallest fee.This can be applied by searching all routes altogether through a connected network (Chelouah & Siarry, 2000).In some applications, it is also helpful to know the shortest paths with two or three additional alternatives, for example, in order to improve the effectiveness of travel information, a driver can get some choices as s/he like based on the need.
Determining routes can use metaheuristic process.It is the method for finding a solution by combining interaction between local searching procedures and higher strategies, in order to create a process that is able to get the local optimal spots, and do a search on space solutions for finding global solutions (Kusumadewi & Purnomo, 2005).Metaheuristic variants are Simulated Annealing, Genetic Algorithms, Cross Entropy, Particle Swarm Optimation and Tabu Search (Bajeh & Abolarinwa, 2011).Searching fastest distance and the shortest route, in this research we choose the tourist attraction places at Rote Ndao Regency-Indonesia is applied using Tabu Search since it has shown good performance with functions more than 10 variables (Sarmady, 2012).This algorithm can optimize costs and can produce minimum cost.In the case of our research, Tabu Search algorithm is able to increase the initial solutions through improving procedures.
As have been stated, we applied the system by examining tourism attraction places in Rote Island, Indonesia.Geographically, it is located between 10 degrees 25 ° -11 degrees south latitude, and 121 degrees 26 49 -123 degrees East longitude.It has a wide choice of tour destinations such as Nemberala Beach, Bo'a Beach, Mando's Beach, Oeseli Beach, Laviti Beach, the Lake of the Dead Sea, Termanu stone, etc.There are many tourists have visited this island.To help tourism development in Indonesia, especially in Rote Island, as well as to examine the application built in finding the shortest route, we conducted this research.

Materials and Methods
This research has been planned for eight months starting from July until November 2018 at the Kupang State Polytechnic, that is to design included the examining the results of the implementation.They were carried out in the Computer Lab.-electrical engineering -Kupang State Polytechnic.The process was based on data taken from information about road route, that related to tourism place, in Rote island.
Rote Islands also called the Island of Roti, is an Island in East Nusa Tenggara Province, Indonesia.East Nusa Tenggara Province itself is located between 80-120 South latitude and 1180-1250 East longitude, with a land area of 47,349.9km 2 and has a total area of a sea of 200,000 km 2 .The number of Islands in East Nusa Tenggara province is 566 Islands.42 Islands have been inhabited while the uninhabited Islands are 524.NTT has got 21 regencies.One of the island, as well as the regency, is Rote (Manafe et al., 2018).
The methods used in this study consists of the following steps.First, we conducted studies on various references libraries related to the research.Topics examined include Tabu Search Algorithm, Graph and the use of Google Maps API.Next, we prepared a database about tourism attractions places in Rote island, which include the name of the tourism objects, etc.After that, we designed application to determine route and distance with the implementation of the Tabu Search algorithm.The last, we examined the application.The steps of our research can be seen in the following figure.Ndoloe, L. A., & Rumlaklak, N. D. (2018)

The Steps of Tabu Search Method
As stated above, we applied tabu search method to determine the shortest route in finding tourism place in Rote Island, Indonesia.The steps of tabu search method are described as follows.

1) Data Collection
The data required for testing algorithms is the data of route and distance as well as tourist sites.

2) Parameters Determination
At this stage, we determined parameters of tabu search, including the number of population of each iteration (N), TabuListMemberCount and itmax_Tabu (maximum iterations of Tabu Search) (See.Munir, 2010).

3) Candidate generation solution of Tabu Search
Candidate generation solution of Tabu Search uses neighborhood selection.Next, we conducted an exchange of neighbor so that the application will appear on the possible 'n' solutions.

4) The selection of the best solution of candidate solutions
Candidate solutions which are generated are evaluated one by one and will be sorted from the smallest to the largest.The best route will be evaluated whether it can be included in the Tabu List or not.In the process, queue function is added in the determination of the neighborhood.

5) Update Tabu List
The best route from step 4 will be evaluated to know whether they are better than what is in the routes of Tabu List or not.If the route appeared is better than existing routes in the Tabu List, then the route will be included in the tabu List, to replace worse routes in the Tabu List.The best route is to be used for candidate generation solutions on the next iteration.However, if the route generated is not better than an existing route in the Tabu List, then the solution will not be included in the Tabu List and random route generation will be conducted for the exchange of neighbors on the next iteration.6) Checking the expiration criteria by considering the maximum iteration specified in step 2.

Start
The shortest route search algorithm with tabu search The Best Travel Route Recommendations

Completed
Select attractions and early departure location In General, the steps described in the method of Tabu Search can be seen in Figure 2 below.

Results and Discussions
Based on the design of the application we have developed, it is explained that the system is built to determine the route of tourism place (Suyanto, 2010).Therefore we listed tourism attractions/tourism places that located in Rote Ndao Regency, Indonesia.These are objects that become tourist destinations.The user of the application is a tourist/ prospective tourists who travel to the attractions of Rote Ndao Regency.The applications can be visited at http://wisata-rote.rrums.compage.The pages of the website are shown in the following Figures.
Figure 7. Main page On the main page, the user selects the attraction to be visited and determine the starting location.There is also additional information available in each option of a tourist destination in the program.The user can also choose "View Description" of the sights she/he wants to see.Description of sights can be seen in Figure 8.After choosing the destination location and starting point, the user has to choose the Next button to continue to the page that will inform the selected attractions and locations of the starting and finishing point.It can be seen in the following figure 9.
Figure 9. Starting /finishing location information and the selection of tour destinations If the information displayed is in accordance with the tourists' choice, then the tourists, in this case, the user, can select the confirmation button to proceed with the searching process.This is the way to get the shortest route algorithm with tabu search.The shortest distance and route page can be seen in Figure 10.It contains information about the best route, the location of each tourist attraction, and distance of tour destinations from the starting point to the finishing point.

Figure
Figure 6.Sequence Diagram

Figure 8 .
Figure 8. Description of Tourism Attractions' Page

Figure 10 .
Figure 10.The shortest route Next, Following are tables that describe data and results of 10 cases with variations of the point, and the number of different destinations.

Table 1
Data tables to the tourism object

Table 2 Table
Data types of Tourism

Table 3
The list of Tourism Places in Rote Island