Adaptive Control of Resource Allocation in a Multichannel Process Using Discrete-Event Simulation

Abstract

Multichannel service and production-support processes route heterogeneous requests among specialized resource groups with limited capacity and partial interchangeability. Static responsibility rules remain transparent but ignore queue state, residual work, class deadlines, and nonstationary arrivals. This paper develops an interpretable discrete-event simulation model for adaptive routing in such processes. The empirical case is the helpdesk of a technical university; the model is stated as a transferable framework for enterprise service and production-support environments with analogous request-routing structures. The dataset contains 4,271 modeled tickets, fifteen request classes, nine independent resource groups, and an expert admissibility matrix with fifty class-group routes. Service times are specified by route-level Beta-PERT distributions and adjusted for group availability. Four routing policies are evaluated under empirical, peak-load, minimum-capacity, and combined stress scenarios. Workload-aware routing reduced mean working-time waiting by 58.66% in the empirical scenario, 57.89% under peak arrivals, 47.30% under minimum capacity, and 73.54% under combined stress. Under the strongest stress scenario, estimated backlog at the arrival horizon decreased by 97.31%. The deadline-risk extension produced only a small additional gain. The results show that transparent state-aware routing can improve robustness without black-box optimization.