EDITOR’S NOTE: More than two decades ago, the foundational elements for successfully utilizing advanced, self-learning algorithms to optimize the procurement process using an agent-based model were well established. Somewhere along the line, we got distracted by SaaS and digital transformation, and now we are at risk of being sidetracked by Agentic AI. One critical point to avoid having this same discussion about failed or failing initiatives 10 to 15 years from now is not to confuse Agentic (equation-based) and Agent (agent-based) development and implementation models.
Member Question:
Have you been able to utilize simulation or optimization modeling to solve a demand-side supply chain problem? If so, what was it?
My Response:
Numerous studies and reports have been published on the various methods (e.g., Monte Carlo) used to determine supply chain optimization.
My preferred method has been the heuristic approach under an agent-based model, in which the unique operating attributes of each stakeholder are understood separately before a collective outcome is identified and achieved.
It is this latter “twist,” if you can call it, that has enabled the optimization process to extend beyond the limitations of execution boundaries referred to in a May 2007 article that appeared in Supply Chain Digest titled Supply Chain Optimization versus Simulation. Specifically, the author’s assertion that “Mathematical” optimality is not used and is not required or likely even feasible. Optimization almost always takes at least some minutes to process (and sometimes hours), and hence isn’t generally usable in an execution environment.”
With partial funding from the Government of Canada’s Scientific Research and Experimental Development (SR&ED) Program, I developed a theory that I refer to as “strand commonality” in which disparate and seemingly unrelated data streams can be linked through the use of advanced algorithms to produce a “positive or beneficial” collective outcome.
It is fascinating that production models have consistently produced the correct results regarding real-world applicability approximately 98.2% of the time.
It gets really interesting when you introduce multiple-tier factors that include intangible elements, such as the value of technician certification versus x number of years of practical experience.
The critical starting point is to recognize that the term supply chain is a misnomer in that it implies a sequential order of events (in the spirit of your question, a non-deterministic set of algorithms that aligns with the equation-based modeling used by most software vendors).
In reality, however, we operate in a world in which synchronization between diverse (and now global) stakeholders has to exchange and quantify disparate data on a real-time basis, and therefore, the term supply practice would be a more appropriate description.
Once again, when you recognize this elemental difference, you will take the first steps toward building an effective and meaningful optimization model.
I hope this helps. In the meantime, I have included a link to the 2007 Supply Chain Digest article.
Reference Links: http://www.scdigest.com/assets/FirstThoughts/07-05-31.php?cid=1073&ctype=content
30
Procurement Insights
Does a 2008 answer to a question on “optimization modeling” still stand in 2024?
Posted on November 4, 2024
0
EDITOR’S NOTE: More than two decades ago, the foundational elements for successfully utilizing advanced, self-learning algorithms to optimize the procurement process using an agent-based model were well established. Somewhere along the line, we got distracted by SaaS and digital transformation, and now we are at risk of being sidetracked by Agentic AI. One critical point to avoid having this same discussion about failed or failing initiatives 10 to 15 years from now is not to confuse Agentic (equation-based) and Agent (agent-based) development and implementation models.
Member Question:
Have you been able to utilize simulation or optimization modeling to solve a demand-side supply chain problem? If so, what was it?
My Response:
Numerous studies and reports have been published on the various methods (e.g., Monte Carlo) used to determine supply chain optimization.
My preferred method has been the heuristic approach under an agent-based model, in which the unique operating attributes of each stakeholder are understood separately before a collective outcome is identified and achieved.
It is this latter “twist,” if you can call it, that has enabled the optimization process to extend beyond the limitations of execution boundaries referred to in a May 2007 article that appeared in Supply Chain Digest titled Supply Chain Optimization versus Simulation. Specifically, the author’s assertion that “Mathematical” optimality is not used and is not required or likely even feasible. Optimization almost always takes at least some minutes to process (and sometimes hours), and hence isn’t generally usable in an execution environment.”
With partial funding from the Government of Canada’s Scientific Research and Experimental Development (SR&ED) Program, I developed a theory that I refer to as “strand commonality” in which disparate and seemingly unrelated data streams can be linked through the use of advanced algorithms to produce a “positive or beneficial” collective outcome.
It is fascinating that production models have consistently produced the correct results regarding real-world applicability approximately 98.2% of the time.
It gets really interesting when you introduce multiple-tier factors that include intangible elements, such as the value of technician certification versus x number of years of practical experience.
The critical starting point is to recognize that the term supply chain is a misnomer in that it implies a sequential order of events (in the spirit of your question, a non-deterministic set of algorithms that aligns with the equation-based modeling used by most software vendors).
In reality, however, we operate in a world in which synchronization between diverse (and now global) stakeholders has to exchange and quantify disparate data on a real-time basis, and therefore, the term supply practice would be a more appropriate description.
Once again, when you recognize this elemental difference, you will take the first steps toward building an effective and meaningful optimization model.
I hope this helps. In the meantime, I have included a link to the 2007 Supply Chain Digest article.
Reference Links: http://www.scdigest.com/assets/FirstThoughts/07-05-31.php?cid=1073&ctype=content
30
Share this:
Related