A hierarchical approach to the sustainable management of a controlled environment life support system

This dissertation examines a hierarchical management approach to the sustainable operation of a Controlled Ecological Life Support System (CELSS). Mass closure requires that system resources be managed in a manner that prevents shortages and/or surpluses in order to avoid violating the sustainability requirements of system components. We express the sustainability of CELSS in terms of constraints imposed on its subsystems. These constraints are of two types: static and operational. Static constraints capture the basic sustainability requirements of the individual subsystem components---they represent the absolute limits on the operational range of these subsystems. Operational constraints, on the other hand, represent a response to global changes in the availability of system resources. They are imposed as the system evolves dynamically to avert shortages or surpluses in various subsystems. This perspective, termed the constraint perspective, accommodates planning at the subsystem level in order to achieve optimality or other operational goals while guaranteeing long-term sustainable operation. It admits simple system representations and leads naturally to a management hierarchy. We investigate the feasibility of this approach to sustainability on abstract, continuous-time, linear CELSS representations comprising a network of compartments and buffers. Concepts from viability theory are used to develop operational constraints at the top level of a two-level management hierarchy. We generalize the method to nonlinear systems with measurement error and model uncertainty.