Decision variables are D and H / D
One is discrete, one is a continuous amount, and the former can determine the range of value of the latter
How is this problem to map to GA, it is indeed a problem, now you can code directly, 1 is directly encoding the pipe diameter, which is of course, but if you mention H / D, but if there is no latter, In theory, it is not possible to optimize (because only the two are only known, I can know I, then the cost function), the other method is not feasible, it uses H / D as discrete Look, there is no problem, but he only took 4 values, this is of course not good, even if it is, the 5 sections need 20 codes (2 ^ 20 = 16 ^ 5)
Now I seem to see how the brothers handle this problem!
Also: D and H / D should determine the optimization of the entire pipe network, of course, the terrain is also a factor, but in the optimization process, the terrain is just an input factor, in other words, in the optimization, a few IF statements can be solved. problem,
It can be simpler: the most complicated situation is: there is a drop, there is a pumping station. Of course, you can set the human-machine interface to ensure flexible, but the fundamental purpose is still to optimize, so, it is better to make the computer to compare itself, and finally give you an optimized program! (For example, in the case of some buried depth, it can be installed. Although the situation can be manually decided, if it is placed in the optimized environment, I believe that the last result is much better, because This is the overall optimization result!)