No single optimization method exists that is best suited to solve all
path generation problems effectively [Erdman 1991]. Graphical illustra
tions of the mechanism design space, given in this paper, attest to th
e difficulties encountered in the optimization of path generation task
s. An example has also been provided to illustrate the importance of a
good initial guess in achieving proper convergence. Use of experiment
al design techniques that are based on statistically designed orthogon
al arrays is suggested in this paper, as an alternate method to solve
mechanism design tasks. It will also be beneficial to use this method
to obtain a good starting point for the traditional optimal synthesis
procedures. Also, when a theoretical solution is translated into physi
cal reality, manufacturing tolerances, joint clearances, and elasticit
y in link members cause variations in mechanism performance. Orthogona
l arrays and the related robust design procedures, suggested here, acc
ount for such variations as well. Design examples are given to illustr
ate the use of orthogonal arrays.