Full Program »
Exploring the Philosophical Implications of Optimal Control in the Sciences (OptiSci) over Vast Length and Time ScalesThe desire to optimize extends across all the sciences. In the laboratory, these activities may range from seeking optimal performance of a quantum-operating device to optimizing the properties of a new material. Similarly, from certain perspectives nature may be recognized as optimizing through evolutionary processes. We refer to these collective efforts to optimize, whether by the hand of a scientist or by nature, as “OptiSci” – Optimal Control in the Sciences. A large body of evidence exists that optimization in the sciences is often considerably easier than a complexity assessment would suggest. This observation has led to characterizing OptiSci formally, by articulating a basic principle, with underlying physical assumptions, whose satisfaction accounts for this striking experimentally-observed control behavior across a broad range of physical and biological systems. Understanding why this formal structure applies so broadly is a central focus of the OptiSci project.
This poster will outline the elements of a collaborative research project funded by the Templeton Foundation involving scientists from the physical and biological sciences and a philosopher of science. It will include three basic parts: (1) an outline of the formal foundation for OptiSci, encapsulated in a set of three basic assumptions, expressed formally, (2) a summary of supporting experimental evidence from quantum mechanics, chemistry and materials science, and natural evolution with micro-organisms, and (3) points of contact between this research and contemporary philosophy of science.
The link between optimization in various sciences as expressed by OptiSci has not been appreciated before, especially the link between the physical sciences and natural evolution. This research offers some surprising results, suggesting that optimization problems often avoid the so-called “curse of dimensionality”; to the contrary, enriching information about a system or considering a larger set of variables can facilitate targeted optimization. How best to explain, or account for these results, is a central question of the OptiSci project. The research makes contact with a variety of issues of current interest to philosophers of science, both methodological and broadly metaphysical. These include the use of optimality landscape representations in the context of control problems, algorithmic techniques for directed search of large, and complex data sets, the application of formal methods in empirical domains, and issues of unification in the context of inter-disciplinarity. More generally, this research provides a forum for reconsideration of “logics of discovery” that have been largely neglected within philosophy of science.
We have three aims in presenting this work to a philosophical audience: (1) to explore the philosophical implications of OptiSci research generally by getting input from philosophers of science regarding both what the philosophical implications are and how best to grapple with them, (2) to consider OptiSci as a potentially unifying framework for scientific efforts across the sciences, one that may help us understand the potential for integration across highly specialized research methods rooted in distinct disciplinary traditions, and (3) to provide an example of how, as a practical matter, scientists and philosophers can join forces to explore scientific research projects as they unfold.
University of Washington
Benjamin Russell Russell
Katharine Moore Tibbetts
Virginia Commonwealth University