Optimum clustering and routing model using CVRP cluster-first, route-second in a 3PL provider

Third party logistic (3PL) providers commonly offer integrated operation, warehousing in addition to the transportation services. A distribution problem experienced by the Logistics & Distribution division in one of the leading 3PL providers in Jakarta is studied in this paper. Starting 2008, this division partitioned Jakarta province into three clustered areas in which a main dedicated courier is allocated for each cluster. The segregation is done based on the house location of these three main couriers. Total of 13 couriers are ready to be dispatched at any time in addition to the three main couriers. This paper consider a variant of the clustered vehicle routing problem (CluVRP) used to determine a new model and solution methodology for clustering, routing and scheduling couriers in a 3PL company. The objective is to determine a set of vehicle routes which minimize the total cost and deliver all packages without violating any vehicle capacity. The application involved two phases. In the first phase, customers are grouped to several delivery areas according to average demand and frequency of demand occurrences. The clustering algorithm used in this phase is the first phase of sweep A variant with goal to balance the load distribution among the main couriers. In the second phase, CVRP for each main courier in each area is solved by comparing Clarke-Wright saving and nearest neighbor methods. The result shows that implementing the new clustering system combined with nearest neighbor method allows the provider to reduce the number of maximum couriers needed to be dispatched daily by 60% on average. It also allows 43% reduction in the total distance travelled by these couriers which leads to minor CO₂ emissions.