Adaptive algorithm

An adaptive algorithm is a computational procedure that adjusts its own operation while it runs, changing its behavior in response to the data it encounters, the computing resources it has available, or conditions in the environment it operates in, rather than following one fixed sequence of steps regardless of circumstances. The adjustment follows a criterion or reward mechanism defined in advance, so the algorithm can favor whichever strategy performs best given what it has observed so far. This makes adaptive algorithms useful wherever the input or operating conditions cannot be fully known ahead of time, since a fixed procedure would perform poorly under some of those conditions while an adaptive one can shift its approach as new information arrives during execution.