Scheduling - Forward of Backward


(George Hicks) #1

We have been working the past year on getting the Epicor Shedule to work for us. So many cleanup things have been done, standards and hundreds of other things still in the works, but using the schedule to some degree. MRP throws a big wrench in things as far as Capacity goes, so that has been our biggest cleanup effort currently.
We are Backward Scheduling in MRP and so the Jobs are Backward Scheduled from the Demand Date (Sales Order Ship By Date) and allow scheduling in the past to avoid bounce.

Our CFO now wants to Forward Schedule and from what I have read and discussed at conferences, it seems that Forward is not that good, and does not take into account as many things as Backward Scheduling. He thinks Forward is a better method of Scheduling, but know little about how Epicor works in Scheduling.

We currently put a Ship By Date on the Sales Order of 4 Weeks out and let MRP Schedule the Job to come up with a Job Start Date. Yes, often the Job Due Date goes past the Sales Order Ship By Date because of issues on methods etc. but the plan is to use the Capable to Promise once we trust the Schedule and use that to come up with a Sales Order Ship By Date instead of just using a standard 4 Week date.

Anyone have comparison experience with both methods? I remember someone at a conference saying that Forward did not work as well so they Global Rescheduled Every Night and locked Jobs they did not want to move. Sounds like too extreme a solution for us, but open to suggestions.

Thanks in advance,
George Hicks

(Josh Owings) #2

This is a great topic. As for forward scheduling not taking into account as much as backward scheduling this maybe true, but my experience has not been in that way. The amount of assumptions from a forward scheduling manner may be more than a backward schedule, but I really do not think so.

A couple of questions for you, are you using finite scheduling? Do you have a finite horizon set? Do you have your production schedule set for the amount of hours available in the work day? Do you have your machines and people in the system? Do you use scheduling priorities?

With the answers it is my opinion that backward schedule with bounce is best. This allows the system to do it best to meet the highest priority demand. The forward schedule will do the same, but it in my opinion it will front end load some items and just get done as fast as possible. So depending on your resource availability and finite schedule settings things could be very similar.

If you backward schedule and have 4 weeks to build and the method takes 3 weeks the system will tell you to start in one week. If you forward schedule the same thing you will finish up one week earlier than expected shipment and depending on the agreement with your customer can you ship early. If not you can’t bill the customer and you have cost spent that you are not receiving revenue on yet. This is the other way I like to explain it, but again I do not think it uses different variables to forward schedule vs. backward. It may?!

(Rob Bucek) #3

OHHH, scheduling conversations…my favorite. I tend to be very opinionated in this area, but forgive me I’ve been scheduling in general for 15 years, and ten years in Epicor in a production control capaicity and it’s my passion so I’m gonna muzzle myself on this forum at least. There are soo many experiences in scheduling, it’s one of manufacturing’s biggest and most complex challenges since it represents so many opportunities for humanity to fail.

First, don’t look to scheduling software to solve your scheduling problems, I don’t care who makes it. It’s just not gonna happen. Epicor’s scheduling system is a lot of things (especially a decade old engine in bad need of an overhaul) but that being said, it’s really halfway decent. In my experience most orgs struggle far before software comes into play and it just amplifies the existing issues.

Whatever shortcomings exist in Epicor’s forward scheduling essentially expose themselves to the backward methodology as well. More often than not we are confusing management practices a million other variable at play in our particular orgs more than pure discipline. The three major factors I see most companies struggling with that lend to scheduling issues are 1 - capacity management (for discussion purposes let’s just lump in forecasting, sales order acceptance, inventory management {insert your excuse ad nausea here}, 2 - execution (stuff breaks, ppl don’t show up, processes are out of control, {insert your excuse ad nausea here}, and lying (lack of inputs to your scheduling system(s) about items 1 & 2).

My point isn’t that Epicor’s scheduling system doesn’t contribute to scheduling challenges, it does, but that it’s so far from the low hanging fruit it isn’t even funny. Once you are really hitting that you’re (I’m generalizing here) either really good or you’re really complex or a bit of both. Getting to one of your original questions, perhaps you should be doing both. Can’t say it enough, scheduling is a complex solution to a complex problem there is no easy button. Every org is different because of so many variables, there is not a cookie cutter solution. However there may be some theoretical approaches to problem solving this issue that can be used across the board. I really need to jot these thoughts down someday.

Sorry about the soap box, I miss this part of the manufacturing experience and need to jump back in it soon at my current org. I’ve spoke about scheduling in general at Insights and many regional EUG meetings over the years. You’re not alone in this struggle, we all struggle with particular aspects of on time delivery, capacity management, WIP control, inventory turn over, efficiencies etc… the list never ends. I’m attaching a presentation i did a few years back, if anything in there resonates with you, or maybe you’ve seen it, or maybe you’ve already checked out long ago not expecting a dissertation but a simple answer, check this box and your world will be rainbows and kittens…oops. I’d tell you to go buy MS Dynamics or NetSuite but I’ll bet you my retirement your scheduling issues wouldn’t go away either. It’s a process problem, not a software problem. Software can make it more gooderer and less gooderer, but it’s not where you should look for answers.

The system does have a great tool set you can leverage for managing your orgs scheduling needs and therein lies it’s true value regardless of the strengths and weaknesses of the scheduling engine. Invest there and you’ll truly mitigate the weaknesses and bring out the strengths. They truly do need to invest some money and time into their operations work horse however, it’s time.

I just needed to type some words i guess. Thanks!

Link to PPTX below…

(Jim Anderson) #4

I was told once by an Epicor consultant that “Forward Scheduling” in Epicor that behind the scenes it really just backwards schedules. So when forward scheduling it calculates the Ship date from a hard start date but that is about it. Once it determines the start date all the operations times are back-filled in a just in time manner. Hope this makes sense. I never verified this but should be easy to do by reviewing one and looking within an operation, if there is any available capacity it will show at the start of the day as opposed to the end of the day. So if your start times within a operation are important then this may be an issue for you.

On another note, we build fiberglass boats and use the configurator. Our line moves use a takt time. We use lean manufacturing techniques. We haven’t come up with a good way to pad these times on the schedule, and We always schedule a boat based on when the boat gels. We try to schedule boats based on truck loads. Depending on the boat there are only a certain combination of boats that fill a truck. i.e. 1 big boat 2 little boats, or 4 little boats. It’s really more complicated than that for example we may only get 2 boats on a truck if one of them has an Arch option. We also have mold rotation and the fact that a specific mold resource stays with the assembly for a number of operations. The other thing that makes it difficult to get good use out of the schedule is our “Start” aka “Geldate” is not the true beginning of the configured assembly. The upholstery kit gets started a week or 2 before the startdate of the boat. So the true start date is when the upholstery kit gets started but that is not how we think of it. Different than an automobile industry the fiberglass boat gets it color when it gels. So it pretty much is locked in the schedule from the gel start date.

I agree with Rob the the tool needs an overhaul to meet the needs of today’s manufacturers. If there was a way to integrate the Takt times for us it would prove to be a more useful tool and if it lent itself well to be a more forward scheduler for companies like ours where the start dates or more critical than the end dates and where you have more of a push process rather than a pull process. To me as a programmer it seems that Forward scheduling in Epicor was more of an after thought. The programmer didn’t want to rewrite the code to do a true forward schedule so that it populated every stage or operation from the start of the day to the finish. On the other hand I am sure it works great for manufacturers that do JIT scheduling as I believe the tool has everything it needs for that type of manufacturer as long as the appropriate buy-in and effort is put in up front.

(George Hicks) #5

Thanks Josh;

I tend to agree with Backward Scheduling and I think I quelled the CEO thirst for Forward Scheduling but explaining that CTP on the Sales Order will do a Forward Schedule to arrive at a date for the Sales Order line, and then it will Backward Schedule from there. This is what I read from the help and seems to be true. So basically the best of both worlds.

To answer your questions:
Yes we Finite.
Horizon set @ 15 Days. (Had it set at 30 but seemed too far out for our needs)
Production Schedule, i.e. Calendars are all set up for every Resource Group.
Resource Groups have # of Resources set up to achieve matching Capacity to Actual shop Capacity. So from a Capacity standpoint, all is well. Working on Crew Size now to get Labor Costing correct, but that is a side project to Schedule.
Yes, we use Scheduling Priorities, and they work as expected.
MRP Jobs are my real hangup at this point, causing overload on Resources for MRP Jobs that we don’t need to firm up.

Great response, much appreciated!