A New Approach to Sequencing Optimization
Optessa MLS, versions 7.0 and beyond introduce multi-zone sequencing, a new approach to optimization which can benefit plants with the need to generate a sequence for multiple production areas—each zone can have its own sequence, with its own set of constraints, and all zones can be optimized at the same time.
> Most modern production does not occur in a single segment, but rather in a series of interconnected steps. For example, in an automotive assembly plant, there may be three key zones—typically, these are Metal, Paint, and Trim. Within a given zone, there may also be one or more lines that need to be sequenced. For example, we may have one Metal stamping shop (M1), two Paint booths (P1 and P2), and then one Trim assembly line (T1). Such a plant topology is pictured below:
Sequencing Optimization Using Decomposition
In the earlier versions of Optessa MLS, we optimized using decomposition. As you can see in the image above, each zone is decoupled and sequenced individually. This means one of two possible things: either (1) customers must sequence each zone individually, optimizing them in multiple phases or (2) customers try to enforce all of their constraints on one particular zone.
> While effective, the two approaches above can have drawbacks. Either, (1) sequencing the zones in multiple phases takes more time for users since they need to import, finalize, run sequence, and export data for each zone or (2) customers applying all of their constraints to one zone, results in the same sequence for all zones. Hence, if the sequences actually differ between the zones, then the other zones cannot be accurately modeled in versions prior to version 7
Introducing the Virtual Flow Line
In order to address these complications, Optessa MLS versions 7.0 and later have a new feature called Virtual Flow Line (VFL). This new feature looks at the bigger picture, linking together all of the sequencing zones and optimizing the zones together. Three key features include:
> Within one model, the position of a given order can be tracked across all sequencing zones. This means that orders are linked across all the zones—and their sequences can differ whether it’s due to buffers, extended lead times, partial routes, or multiple passes through a zone.
> If there are parallel lines, the merge/diverge points between zones are considered.
> The start and end times of each zone are derived from the primary zone, meaning that the time horizon of each zone is automatically calculated and applied during sequencing.
Accounting for Extended Lead Times
Let’s illustrate this by considering the same plant topology from before, but now we will also add extended lead times. This means that certain vehicles in our sequence will need to pass through a paint booth twice. Also, this means that there will be an extended lead time (ELT) associated with the second pass through paint (because the paint from the first pass needs time to dry). Our revised plant topology is pictured below:
> Due to the extended lead times, our two-tone units appear later in the sequence when leaving Paint than when entering Paint, thereby making the sequences at Metal and Trim different.
> With different sequences at each zone, applying constraints becomes more challenging. For example, two units that are back to back in Trim may not actually be back to back in Metal.
Apply Sequencing Constraints Independently By Zone
Using the new VFL functionality, we can model the orders in each of the zones; this allows users to apply their constraints to differing sequences at each zone. At the same time, we can also keep track of both passes through the paint shop—meaning we can achieve optimal color blocking on both passes through the paint booths.
> And the best part, since the VFL logic optimizes all of the zones together, we obtain a solution that is optimized across all the zones as opposed to within an individual zone. For example, not only will our sequence consider the color blocking for both two-tone passes in the paint shop, but it will also consider the trade-offs between color blocking requirements in Paint vs. smoothing requirements in Trim vs. stamping requirements in the Metal zone.
> Optimizing holistically, we can achieve better solution quality across all of the sequencing zones.