Sesh Murthy

A-Teams: An Agent Architecture for Optimization and Decision Support

The effectiveness of an agent architecture is measured by its successful application to real problems. In this paper, we describe an agent architecture, A-Teams, that we have successfully used to develop real-world optimization and decision support applications. In an A-Team, an asynchronous team of agents shares a population of solutions and evolves an optimized set of solutions. Each agent embodies its own algorithm for creating, improving or eliminating a solution. Through sharing of the population of solutions, cooperative behavior between agents emerges and tends to result in better solutions than any one agent could produce. Since agents in an A-Team are autonomous and asynchronous, the architecture is both scalable and robust. In order to make the architecture easier to use and more widely available, we have developed an A-Team class library that provides a foundation for creating A-Team based decision-support systems.

An Agent-Based Approach for Scheduling Multiple Machines

We present a new agent-based solution approach for the problem of scheduling multiple non-identical machines in the face of sequence dependent setups, job machine restrictions, batch size preferences, fixed costs of assigning jobs to machines and downstream considerations. We consider multiple objectives such as minimizing (weighted) earliness and tardiness, and minimizing job-machine assignment costs. We use an agent-based architecture called Asynchronous Team (A-Team), in which each agent encapsulates a different problem solving strategy and agents cooperate by exchanging results. Computational experiments on large instances of real-world scheduling problems show that the results obtained by this approach are significantly better than any single algorithm or the scheduler alone. This approach has been successfully implemented in an industrial scheduling system.

Cooperative Multiobjective Decision Support for the Paper Industry

We built and deployed a decision-support system for scheduling paper manufacturing and distribution, an extremely complex task with multiple stages of production and strong interaction between stages. In contrast to earlier approaches, our system considers multiple scheduling objectives and multiple stages of production and distribution simultaneously using multiple evaluation criteria. Our system functions as an intelligent assistant to the schedulers and generates multiple good scheduling alternatives using a portfolio of algorithms and direct human-expert input. The successful deployment of our system at several paper mills in North America has resulted insignificant savings, greater customer satisfaction, and improved business processes.

Scheduling Solutions for the Paper Industry

This paper describes a decision support system for paper production scheduling. This is the first system to provide an integrated solution to paper production scheduling and to consider interactions between different stages of the manufacturing and distribution process. Using a multicriteria optimization approach, the system generates multiple enterprisewide schedules to reveal tradeoffs between the multiple, often competing, objectives. The large portfolio of algorithms used by the system is embedded in an agent-based decision support framework, called Asynchronous Team (A-Team). Successful implementations of the system in several paper mills in North America have resulted in significant savings and improved customer satisfaction.

Decision Support for Managing an Electronic Supply Chain

The growth of the Internet and electronic commerce opens new venues for companies to create flexible supply chains by offering high-speed communication and tight connectivity. To take full advantage of these new opportunities, companies need to be able to respond quickly to opportunities as they arise and make smart decisions taking into account the vast amount of data now available. We describe a decision support system which helps companies in making decisions about purchasing, manufacturing or promotions, while considering supplies and demands posted on the Internet, in addition to their own inventory, capacity and demand. The decision support system is currently being tested within an electronic marketplace we built for paper products, which went online in September 1999.

Solving constraint satisfaction problems using ATeams

This paper presents an approach to solving constraint satisfaction problems using Asynchronous Teams of autonomous agents (ATeams). The focus for the constraint satisfaction problem is derived from an effort to support spatial layout generation in a conceptual design framework. The constraint specification allows a high-level representation and manipulation of qualitative geometric information. We present a computational technique based on ATeams to instantiate solutions to the constraint satisfaction problem. The technique uses a search for a solution in numerical space. This permits us to handle both qualitative relationships and numerical constraints in a unified framework. We show that simple knowledge, about human spatial reasoning and about the nature of arithmetic operators can be hierarchically encapsulated and exploited efficiently in the search. An example illustrates the generality of the approach for conceptual design. We also present empirical studies that contrast the efficiency of ATeams with a search based on genetic algorithms. Based on these preliminary results, we argue that the ATeams approach elegantly handles arbitrary sets of constraints, is computationally efficient, and hence merits further investigation.