Frequently when we work with clients to implement decision support tools for supply chain scheduling and planning, they often have some unique constraint that is essential to model and may be unique to their environment. Some recent examples we have encountered include the following:
- When producing a batch in a make to order environment, the plant always produces some extra amount called the purge quantity which is stuck in the piping from the reactor to the packout line. After purging the line, this material is recycled into the next batch.
- A warehouse can have capacity constraints on both the
- Throughput based on the number and type of doors and
- Storage based on material characteristics such as hazardous material classifications.
- When working with a dairy industry client, the bill of materials changes throughout the year based on the component ratios of the milk produced by the cows which drives the product split.
These types of situations are a regular occurrence and require modeling tools that allow for the flexibility to deal with them. We will implement decision support tools either with a development suite such as Aspen Tech’s Supply Chain ManagementTM or develop an application that connects to an optimization engine such as FICO’s XpressTM. These tools provide a base starting point but then allow for adding modeling constraints that are required to include to get to a solution that the client can actually implement.
In addition, having this flexibility allows the work processes and tools that enable the work processes to evolve over time as the business needs change.
This flexibility though has to be balanced with some level of standardization. Therefore we will often build a new application by using a previous application as a starting point. For a production scheduling tool, there are many things that are common between different implementations including how to represent the schedule via an interactive Gantt chart, common basic reports, standard interfaces to external systems, etc. In a production planning tool, typically there are plants, warehouses and transshipment points to be modelled via a network representation; costs and capacities at each of these nodes in the network that need to be modelled; and an objective function that is either to minimize cost or maximize profit. All of these would be common elements between different planning model implementations.
- Flexibility allows for
- Modelling essential constraints that may be unique to a particular client’s environment but are required to get to a feasible solution that the client can actually implement.
- Changing the tool over time as the business needs change.
- Standardization allows for
- Faster / cheaper implementation.
- Faster / cheaper support.
- Ease of training when moving to a different role but using similar tools.
Having a hybrid of flexibility with standardization is the best of both worlds!