Simulation assists organizations in assessing the state, performance, anomalies, opportunities and risks of current and proposed supply chains and logistic systems, virtually stress testing their performance and the feasibility of innovations; guiding tactical and real-time decisions by the numerous interacting stakeholders from material and parts supply to delivery of final products and services to customers; and optimizing and redesigning their supply chain and logistics processes and networks. As a professor at ISYE School and Coca-Cola Chair in Material Handling and Distribution at Georgia Institute of Technology, Benoit Montreuil states that simulation in supply chain and logistics requires modeling and replication of numerous distribution and fulfillment centers, local stores, hubs, and factories, which is complex and demanding. Montreuil’s 40 years of experience has made him believe that the challenges in front of the organizations are multifold. In an interview with CIO-Review, he shares his views on how supply chain and logistics could overcome these challenges.
What, according to you, are the challenges that companies face while simulating their supply chain and logistics processes?
End-to-end simulation of supply chain and logistics processes brings with itself a plethora of hurdles, as organizations need to achieve representative granularity while modeling everything from the pick-up and transfer of goods all the way to their delivery, as well as all key planning and operational decisions. This process consumes a large amount of time, requiring from a few minutes to 8-10 hours to run one simulation cycle depending on the case. In addition, the process of gathering, cleaning and entering different data from the various stakeholders of the supply chain and logistics into a simulation model is also cumbersome.
What are the approaches that organizations could adopt in order to eliminate these challenges?
Businesses need to upgrade the virtual modeling of end-to-end supply chain and logistic system components, processes and decision makers in a simulation to ensure that a near-reality digital twin system is fast and seamlessly designed, developed, updated, run, analyzed and acted upon. This ranges from clients, retail stores, online platforms, distribution and logistic facilities, transport service providers, factories and suppliers, including their key operations, resources and decision makers.
Although simulation runtimes can be further reduced by employing multi-core simulation or distributed simulation, these processes are yet to make their way into the simulation of supply chain and logistics.
Now, while inserting disparate data from the organization’s logistics and supply chain components into the simulation process, the time spent on analyzing the data and chartering it, and then testing it before feeding into the simulator dashboard drastically slows down the system. Thus, the industry requires a simple plug-in system that enables rapid data cleaning and prevents having to reinvent the wheel every time.
You spoke about the challenge of applying custom data into the simulation system. Could you elaborate on that?
The complexities associated with feeding custom data into the simulator are very high. Although all the goods, the DCs, and the supplier stores are stable entities, input data for simulating all these separate elements of the supply chain and logistics are, however, distinct. Thus, not having connected databases leads to the concern of interjecting custom data into the system and extends the simulation model elaboration and runtime even more.
Are there any parameters that need to be considered while modeling the multiple elements of the supply chain and logistics, and does it reduce the runtime?
The modeling process incorporates multiple components, suppliers, clients, production planners, routers, and so on which have raised the necessity of crossover agents that can be easily plugged into the simulation system. This way, the recurring role of various components could be simulated in the system by engaging a trained simulation agent to take real-time decisions. Also, the supply chain and logistics industry should focus on plug-and-play software agents that offer decision-making capabilities and also reduce runtime.
Further, organizations can also rely on emulators to reduce the runtime, although their use is not prevalent in the supply chain and logistics. Emulators that utilize analytical approximations could speed up filtering of potential alternatives, limiting the number of simulation runs while providing adequate precision performance results. Another idea is to run simulations of different components of supply chain and logistics on separate simulators, then create simplified yet representative simulation models of these components and finally embedding them into an overall simulation model. This way runtimes and memory requirements will reduce drastically. Such improvements enable operators to embed modeling functionalities for simulation of the impact of advanced technologies across the supply chain, or to realize large-scale wide-scope models of smart hyperconnected logistic systems and supply chains.
What does the future of simulation hold and how can it be further enhanced for the supply chain and logistics?
A lot is happening in the simulation landscape. Numerous industries are working to embed simulation directly into their enterprise and supply chain systems, allowing to run models seamlessly from refreshed data, rather than constantly having to build punctual models for specific projects. Further, to ensure that the users can seamlessly feed their requirements and input data into the simulation system, businesses are moving toward integrating analytics with simulation.
Now, when it comes to enhancing it, I think, the simulation process can acquire a few properties from the aircraft simulators used for training of pilots. The supply chain and logistics departments should also focus on making the simulation visually life-like and participative by taking an inspiration from the gaming industry. Similarly, to embed simulation into a system, we can adapt the level of in-depth capabilities that manufacturers of very-large-scale integration (VLSI) integrated circuits ensue. All of these three businesses are phenomenal at creating visually attractive, life-like 3D images and simulations that are effective in supporting their target users. I would also like to emphasize on the fact that such companies spend numerous millions of dollars yearly to create these highly realistic designs. The supply chain and logistics field needs to follow the same course to upgrade its simulation process, synergistically merging the capabilities of top retailers, e-commerce players, distributors, logistics service providers, manufacturers, consultants and technology providers.
As a professor at a University, what shift would you like to see in the classroom teachings of the simulation process?
I would like my students to focus and indulge more in carrying out experiments on simulation with different input data rather than learning various coding languages like Java and Python, or simulation languages at the core of platforms such as AnyLogic, Arena and Simio. Classes need to become more immersive, participative and collaborative, where students visit the virtual world and observe the designed networks, processes and layouts, different strategies to carry out the simulations using disparate databases, and learn more about the configuration of the supply chain and logistic systems and operations.
Check out: Top Logistics Technology Companies in APAC