Towards The Automated Inference Of Queueing Network Models From High-Precision Location Tracking Data

Abstract

Traditional methods for deriving performance models of customer flow in real-life systems are manual, timeconsuming and prone to human error. This paper proposes an automated four-stage data processing pipeline which takes as input raw high-precision location tracking data and which outputs a queueing network model of customer flow. The pipeline estimates both the structure of the network and the underlying interarrival and service time distributions of its component service centres. We evaluate our method's effectiveness and accuracy in four experimental case studies.

Deriving coloured generalised stochastic petri net performance models from high-precision location tracking data

“…Earlier work conducted by Horng et al [10] is the most closely related to ours. The authors present a methodology designed to infer simple Queueing Network performance models from high-precision location tracking data.…”

“…This place of the PNPM represents the entry/exit point of the underlying system and thus, is actually not part of any service cycle. To avoid such unwanted cycles to be detected by the algorithm we examine the subgraph of G induced by the set of vertices |I − (Server 0, t 4 )(black) = 1 I + (Interm., t 4 )(black) = 1 I − (Server 0, t 4 )(red) = 1 I + (Interm., t 4 )(red) = 1 t 7 |I − (Travel 2, t 7 )(black) = 1 I + (Server 0, t 7 )(black) = 1 I − (Travel 2, t 7 )(black) = 1 I + (Server 0, t 7 )(red) = 1 t 10 |I − (Interm., t 10 )(black) = 1 I + (Rep., t 10 )(black) = 1 I − (Interm., t 10 )(red) = 1 I + (Rep., t 10 )(black) = 1 Figure 6 to enable the accurate representation of the underlying system's customer flow (Figure 7). Interm.…”

Deriving Generalised Stochastic Petri Net Performance Models from High-Precision Location Tracking Data

“…Our work is mostly closely related to earlier work by Horng et al [12]. This work proposed a methodology for inferring simple Queueing Network performance models from highprecision location tracking data.…”

“…Based on the customer's location traces, the first appearance time corresponds to the first timestamp when the customer is identified to be inside the server's service area; the last disappearance time is defined as the last timestamp when the customer is considered to be outside the service area. The customer's exit time is computed in a similar way, by taking the average of the timestamps of the two location updates that correspond to the last appearance and first disappearance [12]. We maintain two time-ordered lists at each service area to store the customers' entry and exit times.…”