Variations in neuro-control strategies with scaling of optimization criteria

An eighth-order nonlinear Hill-based muscle-joint model is used with a global optimization scheme to study normal-to-fast single-joint movement strategies. The optimization criterion includes measures of neuromuscular effort and dynamic performance of the movement. By systematically varying the relative weights of such competing criteria as position error and neural or muscle effort, a continuum of optimized movements is obtained. These range from very fast time-optimal movements which come at high neural and muscle effort, through more efficient fast movements, and then finally to moderate-speed movements coming at much lower cost to any criteria related to effort, efficiency or smoothness. Sensitivity analysis of each individual criterion to the inclusion of the other in the optimization suggests muscle stress to be a particularly effective criterion.<>