Co-design for Multi-subsystem and Vehicle Routing-and-Control Problems

Abstract

Co-design refers to the process of integrating optimization of the physical design with a controller for a system. The challenge in co-design is that the optimization is simultaneously applied to both static/time-invariant (e.g., physical plant design) variables and dynamic/time-variant (e.g., state and control) variables, which can be coupled with each other. The objective of this dissertation is to explore new formulations and approaches in co-design for multi-subsystem and vehicle routing-and-control problems. Specifically, four research questions are considered and resolved. In Research Question 1 (RQ1), the critical issue is how to formulate a class of multi-subsystem co-design problems with convex physical design subproblems and linear quadratic regulator control subproblems, and construct a decentralized solution approach for such problems. In Research Question 2 (RQ2), solution methods for a broader class of multi-subsystem co-design problems than those considered in RQ1 are investigated. In Research Question 3 (RQ3), the question is whether, in the context of co-design, the combined routing and control costs of a fleet of vehicles can be improved if optimal control is introduced into the routing. Finally, an extension of RQ3 is considered in Research Question 4 (RQ4), where the possibility of constructing an integrated vehicle routing-and-control problem with load-dependent dynamics is investigated. Beyond the articles published by the author of this dissertation, the proposed research questions, models and methods presented have not been considered elsewhere in the literature.