Although the physical principles of motion are generally well understood, adequate methods for controlling motion in computer simulations to achieve specified goals or to satisfy general constraints do not exist. The primary objective of this research project is to develop means that will allow scientists to exercise a high level of control over dynamic models. This is achieved by generalizing the formulation, generation, and satisfaction of constraint equations to create an intuitive interface between the user and the computer. An abstract definition of constraints will encompass force constraints, explicit motion requirements, energy minimization, and style considerations. Assemblies of rigid or flexible bodies, representing forms ranging from human figures to mechnical linkages, form the domain of the geometric models studied. The underlying description of the assemblies encompasses both simplified polygonal representations for prototyping and exact B- spline sculptured surfaces to represent mechanical and other figures. Numerical methods found within Finite Element Analysis, particularly the method of weighted residuals, are studied as a basis for permitting interactive teamwork between the scientist and the machine.