Hierarchical Timed Colored Petri-Net Based Modeling and Evaluation of a Bank Credit Monitoring System

Non-performing assets (NPAs) or bad loans limit the banks from lending to new borrowers. As a result, this would slow down the credit utilization that ends up lowering the deposit interest rates and raising the lending rates, thus, discouraging authentic borrowers. This paper presents a Hierarchical Timed Colored Petri Net (HTCPN) based model that simulates an intelligent agents system which automates credit monitoring and follow-ups by relationship managers (RMs) in bank branches. Efficient follow-ups help the RMs track the borrower's repayment and assess the current values of their counter-assets regularly. In case any loan is kept unattended by the responsible RM, an automated reporting escalation is initiated according to an escalation matrix when due checking time is reached. The HTCPN helped model the expected behaviour of the intelligent agents system used for loan monitoring and also allowed to model a scoring system that is used to monitor the performance of the RMs in handling the loans under their authority and the punctuality of their actions. The performance evaluation through cpn tools simulation resulted in performance ranking and scoring of the RMs, and helped collect statistics on the overall performance of the bank that pointed out to the stages and locations where the flaws are more severe thus providing decision support utilities.