Hi Haoming,
Having an understanding topology optimization is good (!), as is has some of the same underpinnings. But Topology optimization is trying to find the (single) 'optimum' solution for the problem you setup. Generative design is not trying to find that (single) 'optimum' solution will generate multiple designs that meet all the criteria that were specified. None of the proposed solutions may be the 'optimum' but all converged solutions should do be sufficient to do the job.
The way I look at it. the topology optimization concept presumes you know all the input conditions and have defined what you believe is success, Reality is typically a lot fuzzier and less well defined, so Generative design has built in algorithms aimed at generating diverging shapes. Then user at that point can evaluate different options presented and dismiss a good number of them because they don't needs that were not defined up front. A design may be functional but would be hard to produce in sufficient volumes, or would cost too much. But you may also now have 3 or 4 designs that work and that are cost competitive and that you can evaluate better.
In other cases a set of variants will not yield any good results because the material selected is simply not suitable. With topology optimization you probably would have reached the same conclusion, but with generative design you can explore multiple materials in one setup.
Chances are because all of this, the design you will end up going with is going to be different than what you would have ended up using in a Topology optimization.
I hope this helps?
Hanno
Hanno van Raalte,
Product Manager - Injection Molding & Moldflow products
