Systems Of Action - A Blue Print For Real-time Decision Support Systems

Abstract

Capability stacks have proven to be a useful tool for organizing and reasoning about business capabilities. At Statoil we have used it for our Integrated Operations initiatives, as well as for other initiatives such as Integrated Environmental Monitoring. In this paper we introduce and illustrate the concept of capabilities stacks, and describe how the concepts of such a capability stack would be implemented using micro-services and autonomous multi-agent systems supporting a real-time decision support system. The introduction of multi-agents changes the system from a “system of record” (where a human operator is firmly in control and the system provides information for the human operator to make decisions) to a “system of action” (where multi-agents analyze and take action based on available information, and where a human operator has less influence (but remains firmly in control) over the system actions. The introduction of micro-services and multi-agents provides numerous benefits from a systems engineering perspective. Micro-services allow for the isolation of capabilities in easily definable contexts that interacts with their environment through a set of well-defined interfaces, while multi-agents allow for simplified reasoning around capabilities through the goals and rules of the agents without necessarily knowing all of the implementation details. In turn, this allows for a useful discussion around capabilities, and by stacking these capabilities according to a few design principles we are better able to understand and reason about the total system capabilities within a reasonable timeframe and with reasonable amount of analysis resources.