Organisations are complex, dynamic, interacting systems, where components influence each other over time with multiple feedback loops and where aggregations of quantities (e.g. backlogs of tasks) are often more important than the individual tasks themselves. For example, in a system describing telecommunications infrastructure, job arrivals, repair times, availability of engineers and seasonal effects all interact with the collective effect of telecommunications faults.
Parameter estimation and models for the structure and dynamics of organisations has up to now been fairly limited to model simple organisational interactions, using models such as DLMs. It is hoped we can expand the existing work in this area to more flexible models which reflect a more realistic interplay between different components of a system, and thus provide better understanding of dynamic behaviour and more accurate system prediction. This will involve the use of appropriate Bayesian inference methodology (e.g. Sequential Monte Carlo, Approximate Bayesian Computation [1, 2]). These inference methods are particularly suitable to models with complex structures and intractable or computationally infeasible calculations, where traditional likelihood-based techniques have difficulties.