After running extend_elevation_data() and assign_inverts(), iterate through the network to minimize the occurrence of negative slopes. This could work by using an "optimization" framework where the magnitude of error from weird slopes etc is minimized by incrementally adjusting other slopes and invert that are unknown.
Trusted slopes and inverts should be held constant, and the rest of the network should be free to adjust within typical design parameters, minimizing "error".
After running
extend_elevation_data()andassign_inverts(), iterate through the network to minimize the occurrence of negative slopes. This could work by using an "optimization" framework where the magnitude of error from weird slopes etc is minimized by incrementally adjusting other slopes and invert that are unknown.Trusted slopes and inverts should be held constant, and the rest of the network should be free to adjust within typical design parameters, minimizing "error".