Proxel-based simulation of queuing systems with attributed customers

This paper describes a state space-based simulation method for queuing systems with attributed customers. The approach extends a previous version, which was designed for only one customer class, enabling the simulation of a larger group of queuing models. The work is motivated by the need for exact solutions for queuing systems where no analytical solution is available. In this paper, the original Proxel-based queuing simulation method is extended to incorporate attributed customers, concentrating on efficient coding and storage strategies to dampen the state space explosion. The attribute classes priority, deadline and processing time are implemented. Experimental results indicate the maximum number of attribute values that is still feasible. Some interesting statistics are presented, which would be hard to obtain using traditional simulation methods. The presented method can yield deterministic results for a larger number of queuing systems that cannot be easily solved analytically.