ANT COLONY OPTIMIZATION MODEL FOR DETERMINING THE SHORTEST ROUTE IN MADURA-INDONESIA TOURISM PLACES

Travel planning is important, especially in areas that often-become tourist destinations. Each region must have an interesting tour, one of which is on the island of Madura. With so many tours available, it confuses tourists in determining tourist routes. In addition, on the island of Madura, many traditional markets spill onto the streets on certain days which can cause traffic jams so that tourists' journeys are hampered. In this study, a research method using Ant Colony Optimization (ACO) is proposed to determine the shortest route to tourist sites on Madura Island. Ant Colony Optimization method is one method that can solve an optimization problem. In solving the problem this method is inspired by the behavior of a collection of ants. Ants function as agents assigned to find solutions to a problem. Based on the experiments carried out, the accuracy value in finding the shortest route solution was 80%. In addition, the number of tours and the magnitude of the distance also affect the execution time of the process of determining the shortest route. The more tours that are visited and the greater the distance traveled, the longer the execution time of the process of determining the shortest route. 2 RACHMAD, SYARIF, ROCHMAN, HUSNI, RAHMATULLAH


INTRODUCTION
Each area has interesting tourist attractions, one of which is tourism on the island of Madura.
Madura Island has various types of tourism objects, ranging from natural, cultural, and religious tourism spread in various regions, namely Bangkalan, Sampang, Pamekasan, and Sumenep. [1] Based on data from the Department of Culture and Tourism for each district on Madura Island, Beach, and so on [2] [3].
Each tourist spot has various potentials, but from these potential problems arise to reach these tourist attractions such as determining the shortest route. To be able to go to tourist destinations, there will be a choice of routes that are passed in each area. In addition to route selection, to reach tourist attractions on Madura Island, many traditional markets spill onto the streets on certain days or commonly called market days which can cause traffic jams so that tourist trips on tours on Madura Island are slightly hampered. From the problems above, a system that supports tourists is needed, namely the shortest tourist route. This route will take tourists to the tourist places to be visited to be able to save time, distance, and cost [4] [5].
Planning a trip before going on a tour is an important thing. With a travel plan, tourists can easily see an overview of the trip and the desired shortest route options to be able to shorten the time to arrive at the tourist destination. In this study, to determine the shortest route to this tourist location using the Ant Colony Optimization (ACO) method. ACO is an algorithm that adopts the behavior of an ant colony. Ants can find the shortest route in finding food sources, this is the nature of ants. It is based on footprints on the trajectory that has been traversed. The more ants that pass through a track, the clearer the footprints will be. 3 ANT COLONY OPTIMIZATION MODEL

Shortest Route
Traveling Salesmen Problem (TSP) is one of the optimization problems such as determining the shortest route [5]. The problem of determining the shortest route can be categorized into optimization problems. Determination of the shortest route is to determine the most optimal path, namely the path with the shortest route and the smallest cost. Time can be related to the distance traveled, the shorter the distance, the shorter the time needed to cover the distance [5] [6].
Calculation of the shortest route plays an important role in everyday life because it must be done in a short time and at the same time so that it can immediately be known which route is the shortest to pass [7].
The optimal value can be found in two ways. The first is the conventional method, which is to try all the possibilities by recording the values obtained, this method is less effective because the optimization will run slowly. The two heuristic methods are using a formula to get the optimal value quickly and precisely. One of the heuristic methods is the Ant Colony Optimization algorithm. This algorithm is one of the best algorithms in solving problems regarding determining the shortest route such as the Traveling Salesman Problem (TSP) [8].

Ant Colony Optimization (ACO)
This optimization problem is solved in several well-known ways, including using the simulated annealing algorithm, genetic algorithm, bee colony, A*, Dijkstra's Algorithm, and Ant Colony Optimization [9]. Ant Colony Optimization or ACO was introduced by Moyson and Manderick and developed by Marco Dorigo. This algorithm is called bioinspired metaheuristic which is included in the Swarm Intelligence group, which is one type of paradigm development used to solve optimization problems where the inspiration used to solve the problem comes from the behavior of the insect swarm. Ant Colony Optimization is generally used to solve discrete optimization problems and complex problems where there are many variables. The results obtained using the ACO algorithm are close to the optimal value [10].
Ant Colony Optimization algorithm is adopted from the behavior of ant colonies in finding food 4 RACHMAD, SYARIF, ROCHMAN, HUSNI, RAHMATULLAH sources. Ant colonies can find the shortest path between the nest and the food source based on the footprints they have passed. The more ants that pass through one track, the clearer the footprints will be. This causes the trajectory that is rarely passed by ants, the density of ants passing through it will decrease or even no one will pass through it. On the other hand, the trajectory traversed by ants in large numbers will increase the density of ants that pass through it or even all ants pass through the trajectory [11].
Ant colonies have a unique behavior when looking for food sources [12]. The ants spread out looking for the shortest path to find the food source. In the process of searching for food sources, ants communicate with each other through a liquid chemical substance left behind as footprints called pheromones. So that other ants will follow and choose a path with a higher pheromone level. This is because the more pheromones in a path, the more ants pass through that path. The pheromone will later experience evaporation, but because many ants pass through this path, the pheromone value will remain strong and the route is the shortest distance traveled. This is what underlies the concept of the Ant Colony Optimization algorithm [13][14].

Accuracy
To calculate the accuracy value using equation (9) which uses the number of correct data compared to the test data [15].

Data Collection
Tourism data on Madura Island is based on data from the Culture and Tourism Office of each Regency in Madura in 2020 as many as 61 tours, namely Bangkalan Regency as many as 21 tours, Sampang Regency as many as 12 tours, Pamekasan Regency as many as 6 tours and Sumenep Regency as many as 22 tours.
Distance data is the distance between tourist sites, distance data is taken from Google Maps, 5 ANT COLONY OPTIMIZATION MODEL traditional market data that causes congestion (on certain market days) is also taken from Google Maps as many as 36 markets, namely Bangkalan district as many as 11 markets, Sampang district as many as 6 markets, district Pamekasan as many as 7 markets, Sumenep district as many as 12 markets

Traditional Market
On the island of Madura, many traditional markets always cause crowds or overcrowding to spill over to the highway axis. In addition, in the Madura traditional market, there is also a daily market or Laki market. This daily market or Laki market sells various animals such as cows and goats. This market occurs once a week with market visitors who are more crowded than other days causing traffic jams that are worse than usual. Thus causing congestion at several points of the highway which has an impact on residents who want to pass by in the area around the market, including tourists who want to travel on the island of Madura. The output generated from the calculation of the shortest route is shown in Figure 1 in the form of maps using Leaflet javascript, Open Street Map, Geolocation, and Leaflet Direction Routing.

Ant Colony Optimization (ACO)
The algorithm used in this research is the Ant Colony Optimization algorithm. The following flowchart of the Ant Colony Optimization algorithm can be seen in Figure 2. a. Matrix initialization dij , dij is the distance from one tour to another in Bangkalan district as shown in Table 1.  Suppose in the 1st tabulist by the 1st ant. Initially, the ant is placed at node 1, then node i is the same as node 1. Then the ant goes to node j.
-Determine node j, Node j is a node between nodes 1 to 50 on the attraction.
The way to determine it is based on the greatest probability value, to find out, formula (1) is used: (1) e. Calculating the Probability on a path with formula (2) (2) f. Counting pheromones while on path is iteration g. Counting pheromone global (4) h. Counting pheromones for each tabulist (5) i. Obtained results such as Table 2 j. Do iterations as many ants The value of k (number of ants) is used when searching for ant-visiting paths. Testing the value of k will be carried out to analyze how the influence of the value of k on the distance results obtained by the ACO method. The value must be greater than 0, therefore the value of k used in the test is 21. The test results for the value of k can be seen in Table 3 and Figure 3.  After getting the results of several k parameters tested, at a k value of 21, the smallest distance value is 188.5 km. Then the value of the parameter k of 21 or as many as tours is the most optimal because it produces the smallest distance value.

Trial of the Ant Colony Optimization model
To determine the accuracy value of Ant Colony Optimization in determining the shortest route to tourist sites on Madura Island, a test has been carried out by comparing the results of the shortest route obtained from the system using Ant Colony Optimization with the possible routes formed. Based on 20 experiments by comparing the distance between the route formed by the system and the route formed by Google Maps, it was found that 16 trials resulted in the optimal route, 3 times the experiment resulted in a non-optimal route, and 1 trial that 12 RACHMAD, SYARIF, ROCHMAN, HUSNI, RAHMATULLAH could not find the route. After analysis, it is possible that the route generated by the system is not optimal because the library from the javascript leaflet and OpenStreetMap can not recognize routes with narrow roads. Unlike Google Maps, which can recognize routes with narrow roads. So that Google Maps can generate accurate routes. Meanwhile, the shortest route was not found due to tours across the island such as Gili Labak, Gili Lyang, and others so that neither the system nor Google Maps could recognize tourist locations, distances, and routes to these tours. This shows the accuracy results using ACO have a success rate of 80%.

CONCLUSION
Based on the research and system testing that has been done, it can be concluded that first, comparative results. Testing the system using ACO in determining tourist routes for all tours in Bangkalan Regency, with the center point at Trunojoyo Madura University without market barriers, resulting in a k parameter value of 21 which is 188.5 km, while the distance traveled if there are traditional market barriers as much as 217.9 km. Second, the system for determining the shortest route to tourist sites on Madura Island using the Ant Colony Optimization method has an accuracy of 80% in finding the shortest route solution. System for determining the shortest route to tourist sites in Madura