Open systems exploration: An example with ecosystems management

Author

Funabashi, Masatoshi

Abstract

Open complex systems in operation can only be managed with internal observation, where the reproducibility and predictability of conventional science with external observation is not necessary valid nor fully achievable. However, globally important problems are mostly open systems that beget one-time-only events. Still, in open complex systems, internal observers may better manage the system by interactive database and interfaces to explore effective variables. We develop plural modalities of such data-driven interface for open systems exploration, taking an example in ecosystems management with citizen observation. The examples are developed from data-supported widening of choice, suggestion with statistical inference, and to a basic setup that can interactively select a best prediction model with inputs on-the-fly. These interfaces were applied to 1-year field observation and yielded a tentative scoring system of index species. We define basic conceptual framework with a realization of the initial steps of open systems exploration, that will subsequently follow interactive reconfiguration as the systems evolve.