Weighted shortest job first
Jump to navigation
Jump to search
…work for one or more quarters of a year.
✣ ✣ ✣
Pick the right stuff to work on to minimuze output (and thus effort) and maximize outcome.
Therefore:
Order the items based on weighted shortest job first to minimuze output and maximize outcome.
✣ ✣ ✣
✣ ✣ ✣
- Task Sequencing
- FIFO (First In First Out) ≡ FCFS (First Come First Served) [yet, they are different: First Served does not guarantee First Out]
- FIFO is the right solution in a homogenous cost of delay and task duration. [so, pearl language can be used here: context determines which solution yields best results]
- In Development, cost of delay and duration are non-homogeneous. What to do?
- ¶ high profit first (HPF)
- ¶ minimum slack time first (MSTF)
- {{p|high delay cost first (HDCF)
- {{p|shortest job first (SJF)
- weighted shortest job first
- E.g. run an Emergency Rescue in a hospital, really would like to use some of these TPS practices to my advantage, so I will run FIFO on my incoming patients… “You're insane!” would be the appropriate reaction.
- Some patients have higher urgency. Some block the OR for a very long time. Take Cost of Delay (= Urgency) into account.
- In Operating System Design we have already dealt with these problems.
- Low Hanging Fruit (a.k.a. High ROI) does not equal high Cost of Delay.
- It is silly to favor a 10k/m CoD over a 10M/m CoD because of it is having the most depleted buffer depletion. It optimizes the least amount of complaining customers. You get the most number of jobs done with the amount of resources available.
- ¶ high delay cost first makes more sense since you're going to delay the one with the highest cost of delay. 85% of developers are clueless about their CoD.
- ¶ shortest job first makes sense, too, when everything has the same CoD, which one are you going to pick next? Right, the shortest job first. That way you avoid the convoy effect of one big job blocking a lot of the small jobs.
- To revert back to FIFO would be a giant step backwards.
- ¶ weighted short job first is probably the best technique—weight the economic job by the cost of delay