Dynamic pricing mechanism for multi-agent based system of well scheduling

摘要

Well scheduling is a complex business process which includes the tasks of planning and execution of locating land, constructing, drilling, completing, and tie-in of a well. In our previous work, a multi-agent based framework was proposed to automate well scheduling, where a challenge was to automate the estimation of the price for the tasks that the oil and gas industries may pay and the service providers may accept while gaining economic efficiencies. In this paper, we propose a mechanism to automate this negotiation process. When a new well is added, a scheduling agent (SA) releases the tasks in bundles, and a set of service provider agents (SPAs) bid on the bundles to achieve them. For bundle creation, combinatorial auction is used by the SA, whereas for bid forecasting, a time-series analysis of the historical data is performed by the SPAs using auto regression. The results are discussed to ensure wide acceptance of this mechanism in well scheduling.