When FORCE_STARTJOB is used to start a File Watcher job, the next start times of the FW job are re-calculated at every scheduled interval while the FW job is already RUNNING when it is force started.
The following message is displayed in the EP log:
[01/02/2009 20:25:00] CAUAJM_I_40245 EVENT: STARTJOB JOB: PDWHDMST0030[01/02/2009 20:25:00] <Event was scheduled based on Job Definition.>CAUAJM_I_40136 JOB: PDWHDMST0030 Current STATUS: RUNNING - Cannot be Started.
According to the job definition, AutoSys will try to schedule the job but if it is already running, AutoSys scheduler will not start the job.
In case you define the job this way (to achieve the file is being monitored every x minutes), then you can give the success condition for that job as the success of the same job.
Doing so will ensure that the job is continuously monitored. This will only apply if you FORCE START the job for its first run; otherwise it won't start because the starting condition has not been met.
Same behavior exists with Autosys 4.5.