Optimization of Tour Scheduling Using Firefly Algorithm

Bandung is a region in Indonesia that has good potential tourism. Every year, the number of tourists coming to Bandung is increasing. However, most tourists do not have enough information to organize a tour for visiting several touristic places. Therefor a system for tour scheduling is needed to help tourist to enjoy their holiday. Here, we develop a tour scheduling system using the Firefly Algorithm. This algorithm generally is used to solve Traveling Salesman Problem (TSP). The Firefly algorithm generates tour scheduling based on three criteria of user needs, such as distance, popularity and budget. We calculate these three criteria with the Multi Attribute Utility Theory (MAUT). Each criterion has a priority value from user. We use the result of MAUT as a fitness of Firefly Algorithm. The Firefly algorithm also considers opening hours and closing hours for each tourist attractions. This system also recommends some hotels based on two criteria, i.e., distance and facilities. We calculate these two criteria using MAUT. For evaluation, we compare n-days tour scheduling modeled by one step optimization, with per-day tour optimization (n-steps). We employ some parameters, i.e., running time, fitness value and number of visited destinations. Based on the experiment, the running time of per-day tour optimization is 7% faster than one-step optimization. Meanwhile, the fitness value of per-day tour optimization is 14% higher than one-step optimization. In addition, the number of nodes in tour of per-day tour optimization is 30% more than one step optimization. Moreover, based on popularity, travel time and budget, the higher fitness of per-day tour optimization is obtained.