Carbon Footprint Analysis in Logistics Using Machine Learning Approaches for Sustainable Transportation

Abstract

Climate change has become a significant concern for humanity in this ethnocentric era.. Transportation contributes approximately 23% of the annual global CO2. Governments and politicians have been asked to take action to reduce CO2 emissions in several sectors over the past 20 years. Many nations are currently putting their attention toward lowering CO2 emissions and adopting the development of low-carbon communities. The decrease of CO2 emissions from transportation is one aspect of low-carbon city development. In order to achieve low carbon emissions, studied variables affecting CO2 emissions and proposed methods through which we can reduce CO2 emissions. However, reinforcement learning (RL) has received little attention as a potential solution to the ”traveling salesman problem.” In order to reduce CO2 emissions, this research uses RL, the travelling salesman problem is solved, and an optimal route is provided.. The distances acquired using the RL method and the farthest insertion algorithm were compared. When compared to the farthest insertion algorithm, distances calculated employing RL approaches were more optimal. The optimum distance was obtained using the farthest insertion algorithm and the Q-Learning algorithm, which reduced CO2 emissions. Depending on the mode of transportation, courier transport CO2 emissions vary greatly. Findings indicate a relationship between fuel use and vehicle type and CO2 emissions. Fuel consumption will increase CO2 emissions at a faster rate. CO2 emissions will be decreased by travelling at an optimized distance.