Statistical Analysis and Optimization of Business Load Balancing of a Bank Branch Service Network

• Activity Number: 355
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics and Marketing
• Author(s): Bin Zhang*+ and Ming Xie and Jin Sun and Yuhui Fu and Jin Dong
• Companies: IBM Research China and IBM Research China and IBM Research China and IBM Research China and IBM Research China
• Keywords: business load balancing ; integer programming ; ILOG CPLEX ; optimization

Abstract

A bank branch usually maintains a service network which contains several subsidiary service nodes at different places, and has a number of clerks working for that network. Developing a suitable scheduling table, i.e., determining for each clerk which days he should work and which node he should work at, can be a tough work if only by rule of thumb, especially when the number of nodes and clerks are large. Through statistical analysis on the data from some bank branch, we found that the current scheduling table of the bank branch is not reasonable in the sense that it may result in imbalanced business load distribution, e.g., the average business load per clerk differs greatly in different nodes in the same day, or in different days in the same node. To reduce the imbalance of workload, we developed an integer programming model of scheduling, taking into account realistic constraints and objectives raised by the bank branch. We used IBM ILOG CPLEX to solve this problem and obtained a much balanced scheduling table.