Multilevel Feedback Queue
The term multilevel feedback queue called a dynamic priority scheduling algorithm. In this process there are more than one queue (English queue ) with different priority. Processes are regarded as a function of their previous consumption of resources dynamically in one of these queues.
Realization
There are several uses FIFO queues. The execution is as follows:
In the multilevel feedback queue method, a process has only one chance to be completely processed in a particular queue before it is pushed into a lower queue.
The scheduler always tells the process at the beginning of the topmost non-empty queue to the processor.
Pros and Cons
Benefits
- Short jobs are preferred because they are assigned higher priorities.
- Jobs with many input and output operations are preferred because they are classified in the original waiting list for a voluntary contribution of the processor again, so retain their priority.
- New processes are classified quickly and classified into a priority class. It requires no heuristics (eg, to estimate the duration of the process ) for the classification.
Disadvantages
- If always meet new processes, compute-intensive processes can " starve " the lowest priority class (English starvation ), ie they are never executed until the end.
- Priority inversion can occur.