#20220111 Package Delivery System using Artificial Intelligence and Reinforcement Learning

Within the scope of this project, packages under certain conditions should be able to be transported between specific destination points by an intelligent agent. It is planned to design a faster and more environmentally friendly system where less labor and cost are at the forefront. With the methods to be used and the goal of achieving the best, packages delivery agents can be trained for good behavior. In this project, reinforcement learning is a sub-branch of machine learning that works interactively with different fields such as psychology and neuroscience. Unity Engine tool is used to perform simulation tests close to real-life applications. In line with these purposes, a project that can be implemented in different areas in real life is desired to be implemented. As the focal point, the package delivery processes used in the industry and warehouses are prioritized. Reinforcement learning, unlike other sub-branches of machine learning, determines optimal actions thanks to the interaction of the artificial intelligence (agent) with its environment. There is a dynamic environment. The primary purpose of artificial intelligence is to earn the maximum reward as punishment or reward according to the actions taken. The most critical challenge in this project is to apply the concepts of exploration and operation.

Within the scope of this project, packages under certain conditions should be able to be transported between specific destination points by an intelligent agent. It is planned to design a faster and more environmentally friendly system where less labor and cost are at the forefront. With the methods to be used and the goal of achieving the best, packages delivery agents can be trained for good behavior.

By using Reinforcement Learning methods, the agent's packet transport is provided. With artificial intelligence taking place in our lives every day, researches on this subject are important. Many automation robots run the commands given to them in a sequential manner. Reinforcement Learning agents provide ease of use to the user by learning the task to be done by themselves.

The materials necessary for the project are available today.

-

-

With the completion of the project, the role of carrier robots in the production network will increase and the use of artificial intelligence in robotics will be tested. Reinforcement Learning type artificial intelligence has the potential to achieve success faster and faster than hand-written artificial intelligence, especially in complex transport problems, by optimizing tasks with the machine learning method. With this project, both reinforcement learning type artificial intelligence and the use of artificial intelligence in automation are expected to increase.

-