Toward cache based decision making embedded control systems
Dominika Swierczynska, Tomasz Stach, Mariusz Pelc · 2017
The increased interest in autonomic system requires that control algorithms must facilitate implementation of some kind of decision-making process. Such a process may be to a different extent supported by a human operator who will analyse the system behaviour and, if needed, will make amendments to a control strategy. However, a human operator can only process a limited amount of information at a time. Hence, one can see an increasing interest in various kind of self-learning technologies that would support an increased autonomy of control systems so that the major part of the control, including adjustments to the control strategy, was fully automated. Typically, the deployment platform that is responsible for processing the information about a controlled system and provide control decisions are various kinds of embedded systems. These systems are known of increasing processing power yet the resource constraints are still a problem, especially when it comes to processing huge amounts of environmental information in real time. Implementing self-learning capability apart from an up-datable knowledge base will also a decision(s) cache that would allow to provide some remembered decisions rather than evaluate them from the very beginning. The aim of this paper is to analyse impact of a database system on the performance of the embedded control system through running sets of benchmarks involving various database systems.