Rule-based optimization simulation

Pure simulation approaches are proposed by Pitty et al. (2008) and Adhitya and Srini-vasan (2010). Pitty et al. (2008) propose a discrete-event simulation model for a refinery supply chain. Operational decisions such as unloading schedules and production planning are made based on simple priority rules. Various configurations of the modelled SC are studied and compared to reveal optimization potentials. This approach explicitly considers some details of ship and pipeline transports. Adhitya and Srinivasan (2010) describe a discrete-event simulation model for an SC producing and distributing lubricant additives. Here, batch production is modelled. Again, operational production decisions are made by priority rules and a scenario analysis is conducted to evaluate the effects of other priority... 

We established the scale-up rules to design DFC burners, mainly using numerical simulation. We optimized the scale-up rules for the ejecting velocities and their nozzles arrangement of the raw material and each gas. We also optimized the rules for the distance between the burner and the surface of the molten glass. On the bases of these optimized rules for scale-up, we designed and manufactured a pilot-scale DFC burner. 

The genetic algorithm is a probabilistic optimization method, based on the rules of iterative selection. The method was developed during the 1960 s and 1970 s based on the academic studies of John Holland, but became popular as a practical simulation tool after David Goldberg implemented the academic ideas of evolutionary computation to solve complex applied problems. In the following, the basic ideas and the main steps for the implementation of the method are summarized. 

---