Enhancing market service and enterprise operations through a large-scale GIS-based distribution system

• Propose a two-stage model to solve a large-scale vehicle routing problem.
• Define a GIS-based network.
• Partition distribution area using GIS data and modified K-means algorithm.
• Use hybrid heuristic algorithm to solve routing problems for each cluster.
• The proposed system impacts financial performance of the studied firm.

We develop a large-scale service information system integrating market analysis, customer relationship management (CRM), and distribution service optimization. The system is based on geographical information systems (GIS), and the goals include reducing the distribution cost, increasing the efficiency, satisfying customer demand, and improving service quality. We discuss in detail the design of the model and the implementation of the system. The main contributions of this work are: (1) proposing a new workload evaluation method based on statistical analysis of the large data set. The workload measure is based on GIS and enterprise databases, which address the workload imbalance issue in distribution; (2) implementing an optimal distribution model to serve nearly a hundred thousand retailers. The model contains two stages and uses a Cluster-First-Route-Second approach. In the clustering stage, we improve the K-means method; while at the routing stage, we design a hybrid heuristic algorithm for GIS data by employing the genetic algorithm and simulated annealing techniques; and (3) integrating market analysis, CRM, and distribution service optimization to improve market service and enterprise operations.