aysusayin / fiting-tree Goto Github PK

In the shrinking cone algorithm, we calculate the lowest high slope and highest low slope for each segment. This defines the range of feasible values for the slope of the segment. However, which value should we choose? Taking the average seems like a very straightforward solution but can we think about something else to minimize the error in the access time?

It might be helpful to read that part of the paper again.

When the buffer is full, it needs to be segmented again. How to allocate a new buffer in the segment during this process? Algorithm1 did not talk about allocate a new buffer：