Manufacturing is a challenging real-world domain for studying hierarchical MDP-based optimization algorithms. We have recently obtained very promising results using a hierarchical reinforcement learning based optimization algorithm for a 12-machine transfer line. Transfer lines model factory processes in automobile and many other product assembly plants. Unlike domains such as elevator scheduli...

We present a class of optimization strategies that can be applied to a variety of complex engineering design problems that have an exploitable hierarchical structure. In addition, we demonstrate the power of these strategies by applying them to the design of nanoporous materials, which hold promise in many applications, including those in energy production and storage. In the hierarchical syste...

There are four primary goals of this dissertation. First, design a competent optimization algorithm capable of learning and exploiting appropriate problem decomposition by sampling and evaluating candidate solutions. Second, extend the proposed algorithm to enable the use of hierarchical decomposition as opposed to decomposition on only a single level. Third, design a class of difficult hierarc...

Conversational agents that use reinforcement learning for policy optimization in large domains often face the problem of limited scalability. This problem can be addressed either by using function approximation techniques that estimate an approximate true value function, or by using a hierarchical decomposition of a learning task into subtasks. In this paper, we present a novel approach for dia...

Multi-level TAM optimization is necessary for modular testing of hierarchical SOCs that contain older-generation SOCs as embedded cores. We present two hierarchical TAM optimization flows that exploit recent advances in TAM design for flattened SOC hierarchies. Unlike prior methods that assume flat test hierarchies, the proposed methods are directly applicable to real-world design transfer mode...

This paper is concerned with decentralized controller design for large-scale interconnected systems of pseudo-hierarchical structure. Given such a system, one can use existing techniques to design a decentralized controller for the reference hierarchical model, obtained by eliminating certain weak interconnections of the original system. Although this indirect controller design is appealing as ...

In this paper we deal with the dual criteria optimization, explore the complexity status of hierarchical scheduling on a single machine. Hierarchical optimization is a special case of dual criteria optimization, where the most important criterion is at its optimal value. We'll assume the complexity of hierarchical scheduling on a single machine. While doing this, we'll use the case of two crite...