Set-based Fitness Comparisons - Could Neuroscientists Benefit from Engineering Studies on Conceptual Design?

Since decades ago, biologists have recognized the existence of tradeoffs in biological systems (e.g., [1]). Yet, the first study that borrowed the idea of Pareto-optimality to analyze tradeoffs in biological systems was published in 1995 [2]. In 2008, following [2], it was hypothesized in [3] that Pareto-optimality, which has been used by engineers for many years, should prove to also be significant to the understanding of natural systems. It was also argued in [3] that terminologies such as multi-objective optimization, which are associated with the Pareto principle, should be carefully revised to fit into our understanding of natural adaptation. Yet, it was probably the study in [4] that set the stage for the increasing use of Pareto-optimality by researchers on biological systems, including neuroscience researchers (e.g., [5]). Given the current numerous pieces of evidence for the suggestions in [3], it is argued here that a related issue raised in [3] should also be examined. Namely, could biologists benefit from studies on set-based conceptual designs? Here, this question will be addressed concerning neuroscience. Research on set-based conceptual design commonly deals with multi-concept multi-objective search and optimization [6]. In such studies, a conceptual solution (concept) is viewed as a solution idea that is represented by a set of particular designs that share some common features. While many of the reviewed studies in [6] have applied Pareto domination relation to support the selection of a concept, it was argued in [3, 7], that concepts should not be compared by vector-based domination, as in Pareto-optimality, but rather by set-based domination relations. Following [3] and the existing studies on set-based conceptual design, as reviewed in [6], it is argued here that neuroscience research could benefit from using the mathematical concept of set-based domination, rather than applying just vector-based domination. First, in this presentation, several set-based domination relations will be presented and discussed viz-a-viz existing Pareto-based neuroscience studies. Next, several ideas on the use of such relations for neuroscience research will be highlighted and illustrated.