A survey of approximation algorithms for capacitated vehicle routing problems

Finding the shortest travelling tour of vehicles with capacity k from the depot to the customers is called the Capacity vehicle routing problem (CVRP). CVRP plays an essential position in logistics systems, and it is the most intensively studied problem in combinatorial optimization. In complexity, CVRP with k $\ge$ 3 is an NP-hard problem, and it is APX-hard as well. We already knew that it could not be approximated in metric space. Moreover, it is the first problem resisting Arora's famous approximation framework. So, whether there is, a polynomial-time (1+$\epsilon$)-approximation for the Euclidean CVRP for any $\epsilon>0$ is still an open problem. This paper will summarize the research progress from history to up-to-date developments. The survey will be updated periodically.