Improved Genetic Algorithm and Mobile Application for an Up-to-date Traveling Salesman Problem

Nowadays, with the increase of e-commerce sites, the cargo sector has been growing rapidly. Also, the pandemic process in consequence of COVID-19 virus in the world shows that cargo transportation is getting more important. In only Turkey, daily distance access reaches 5 million km and couriers visit 7.5 million addresses in one day averagely. In addition, cargo companies are competing to maximize their profits and make more deliveries. In this study, Travelling Salesman Problem (TSP) has been focused on. This problem is about searching the optimum route containing lots of destinations. The route calculations in literature are implemented in various ways by using the machine learning algorithm. Mostly, genetic algorithms are encountered as solutions for TSP. In this study, in contrast to traditional genetic algorithms, a novel genetic algorithm supporting multi parameters towards the requirements of the cargo firms is proposed. Thus, six options in routing calculation have been determined and provided to address the needs of carriers. These selections are the only distance, only duration, both distance and duration, only distance and customer priority, only duration and customer priority, and all of distance, duration, and customer priority. According to the selection, genetic algorithm parameters are set to calculate routes. In this way, "customer priority, if necessary, the fastest distribution or the most savings" can be provided. Moreover, it is defended that saving time, increasing the profitability rate of cargo companies, increasing the satisfaction of users and customers, furthermore, reduced carbon emissions indirectly can be provided accompanied by the study.