Combinatorial optimization challenge: delivery route planning optimization
Combinatorial optimization is a very important subfield of computer science, which aims to find the optimal solution under a series of constraints. Many practical problems can be classified as combinatorial optimization problems, such as travelling salesman problem, packing problem, etc.
Most of combinatorial optimization problems are difficult to solve and require extremely long running time. The traditional solutions are mainly based on searching algorithms. With the rapid development of machine learning and deep learning, especially deep neural networks and deep reinforcement learning, more efficient solutions can be provided for solving combinatorial optimization problems.
In telecommunication network, combinatorial optimization also plays a key role. One of the typical problems, which is also our challenge this time, is the global optimization of network resources. By optimizing network resources, it is possible to provide more users with higher-speed and more stable network services. In order to help understanding the problem, we have omitted the complicated telecommunication technology and abstracted it as a delivery route planning problem. On the dataset we provide, the solution that gives the lowest cost with minimum time will be the winner.