First come first served is a simple scheduler, where any new tasks are placed at the end of the queue, and the CPU works through the queued tasks in order. This is simple, so very little CPU time is used in scheduling the tasks, maximising the time it can use to complete the tasks.
In round robin scheduling, all new tasks are added to the end of the list, much like first come first served. However, the CPU does not work on each task until it is done - if a task takes more than a set number of CPU cycles, it is sent to the back of the list, so the CPU completes each task bit by bit. You can alter the max cycles in the controls above, so why not see how it works?
In shortest time remaining scheduling, the tasks are sorted in order of the amount of clock cycles they are estimated to take. Any new tasks are inserted into the appropriate position on the list, and the CPU always works on the shortest task available. The downside of this is that it means longer processes can often face 'starvation', where they are persistently denied CPU resources or clock time and are never finished.
Similar to shortest time remaining, shortest job first sorts the list and prioritises shorter tasks that are easier to complete over longer, more demanding tasks. The difference is in how it treats partially completed tasks, which have been paused to complete an interrupt or for another reason. Unlike shortest time remaining, shortest job first will not take into account partially complete tasks, ranking them according to their estimated complete time from start to finish.
Multi level feedback queues are a complex form of scheduling, that are considered to give the best results. Multiple different queues, each with a different priority rating, are simultaneously maintained. The CPU continually processes the tasks in the highest priority queue that has tasks in, and anything that takes too much CPU time is moved down a priority level (unless it is at priority 0) and is pushed to the back of the queue. The time limits for each queue can be adjusted using the controls above. Processes that are ignored for too long and facing 'starvation' are increased in priority and pushed to the back of the queue (if not already at priority 5), so they can be dealt with sooner. The wait limit before tasks are increased in priority can also be changed using the controls.