If you are manning a data centre with high probabilities of voluminous transactions (heavy OLTP apps), you would wish to take backup of logical logs in continuous manner, than frequency based execution. When a continuous backup is fired (eg. ontape -c or onconfig based ALARMPROGRAM specification via log_full.sh using onbar) a log is backed-up as soon as it is filled-up, this way, you need not monitor the situation in particular and there is no need to have a dedicated backup operator as well.
If your instance have less transaction activities, and if you intend to schedule backup job on every 15 minutes or so, it might happen that the backup utility may not have enough data to write on each such frequency, defeating the very purpose of scheduling such job. If so, it would be better to have an "on demand" logical log backup. A scheduler (eg. cron) present in an OS itself has an overhead on resources of a server to manage job submitted to it.