Abstract
Asset management teams face many challenges with increasing asset complexity, increasing data volumes, and staff in high demand while optimal asset performance remains paramount. More effective development and management of hydrocarbon assets may be achieved through an increased level of automation of work processes and decision-support technologies across the upstream value chain. Asset management workflow automation poses special challenges as such workflows are multi-disciplinary, cross-functional, and human expertise-intensive. Advancements in the fields of artificial intelligence, software engineering, and decision sciences have led to the development of Multi-Agent Systems (MAS) which have been the cornerstone of achieving higher levels of system and work process automation and autonomy. Here, the human expertise, obtained via knowledge elicitation of domain experts, is encoded into the software in an extensible and sustainable way. Each agent functions as an entity of a distributed computing system, performing a broad range of tasks which may include advanced analytics, data quality checks, and oversight of other agents. Although agents may have competing priorities, they can convey information to one another to broker a feasible decision through collaborative and coordinated decision making. This approach is useful for distributed real-time monitoring and capable of providing technical recommendations under open-loop environments and process control in closed-loop environments. This paper provides insights into a multi-agent development process for hydrocarbon asset management workflow automation and decision support.
