The problem of inverse optimization is to find the objective function that is being minimized, given knowledgeof the constraints and observations of local minima. In this thesis, we consider the special case in whichthe objective function is a linear combination of known basis functions weighted by unknown parameters.Therefore the aim is to recover the weights governing the objective function. We propose a solution approachin this case that is based on the application of necessary conditions for optimality. We begin with a review ofhow these necessary conditions arise, with a particular focus on the relationship between duality theory andinverse optimization. We then proceed to describe our solution approach. Finally, we apply our approachto find a model of goal-directed human walking from experimental data with human subjects.
【 预 览 】
附件列表
Files
Size
Format
View
Inverse optimization of discrete-time systems applied to human locomotion