Some Principles of an Expert Optimization of Assembly

Since the part design process is mainly based on constrains like geometry and function, a static optimization approach can be used. Corresponding optimizing algorithms would yield appropriate values for the part and its function. A more complex situation takes place in the manufacturing and assembly processes: their complex functions are time dependant and may run in a variety of logical arrays and use storage (puffer). Furthermore, in such production branches like automotive, aircraft, plant engineering there are not only machines but also workers who are responsible for the final product assembly. The work force determines the manufacturing process quality in many ways, whereas the actual behavior of working force is much more unpredictable. Such an uncertainty applies as well during the normal operation as in a trouble situation. The paper registers some optimizing methods but puts emphasis on expert optimization where conventional optimization methods are paired with simulations of worker actions in virtual worlds. The assembly cells would be built in virtual environment before the „job one“ schedule and the process would be assessed by experts. Respective conditions to such an approach will be discussed based on a final assembly process in automotive manufacturing.