In this paper, we present an intelligent agent based framework for scheduling and control of an automated warehouse. Order trays and gantry robots (picking devices) are modeled as intelligent agents that behave in a cooperative manner so as to achieve personal and system wide goals. The model has been applied to a real world warehouse using actual data. The system is capable of significantly shortening picking cycle time without compromising on system efficiency. It also produces significantly fewer picking errors compared to a hierarchical model used by the company.