Abstract:

Concurrent banking transactions, a necessity in modern digital financial services, are highly vulnerable to deadlocks when multiple customers compete for shared fund pools. This contention leads to system freezes, data inconsistencies, and potential financial losses. This paper introduces the Banking Simulation System (BSS), a novel and scalable adaptation of Dijkstra’s Banker’s Algorithm designed to proactively ensure deadlock-free fund allocation in high-concurrency banking environments. The BSS models customer requests as concurrent processes and bank fund pools (e.g., checking, savings) as allocatable resources. A critical adaptation is the enforcement of credit discipline by capping a customer's maximum loan claim at 60% of their verified income certificate. For each new transaction, the BSS uses standard resource matrices (Allocation, Max Need, Available) to check for safety in real time.A request is only approved if the new state is still safe, which means that all current transactions can finish without having to wait in a circle.