Business processes need to adapt to changes in the operating conditions and to meet the service-level agreements (SLAs) with a minimum of resources. Changes in operating conditions include hardware and software failures, load variation and variations in user interaction with the system. An integral component to adaptation is the awareness over the behavior of self and environment (or having an estimation of the current situation). It is also desirable that businesses have the ability to forecast business metrics and key performance indicators (KPI) to prevent service level agreement violations, penalties and to also keep customers satisfied. In a continuously dynamic environment, forecasting models can provide reasonable accurate predictions for effective planning and dynamic decision making. We investigate the automatic building of a dynamic predictive model of the business process that is used for business process optimization. The model is a simulation model whose parameters are tuned at run time by tracking the system with a particle filter. We use forecasting techniques on past system's state estimates generated sequentially over time to automatically identify, analyze and forecast the behavior of the system. |