Overview, Project of Dr. Hossein

Supply chain network design is the strategic design of the physical and decision structure through which products move from supply sources to end customers. In practice, it addresses questions such as which suppliers, plants, warehouses, distribution centers, or hubs should be used, how the network footprint should be configured, which customer markets should be served from which facilities, and how material and product flows should move across the network to meet service requirements at the lowest practical total landed cost.

Supply chain optimization is broader. It focuses on improving the performance of the end-to-end supply network so that sourcing, production, transportation, inventory, and fulfillment decisions work together efficiently and responsively. In business terms, the main questions typically include what footprint should be operated, which facilities should serve which markets, how inbound flows should move from suppliers to plants or distribution centers, how outbound flows should move from plants or distribution centers to customers, what service levels should be targeted, and how total cost, profit, responsiveness, and resilience should be balanced.

At the most basic level, many enterprises face a strategic network design problem centered on facility location, footprint configuration, and flow allocation across the network. This includes decisions such as where facilities should be located or retained, how nodes should be assigned to upstream or downstream facilities, whether the network should follow direct-shipment patterns or hub-and-spoke structures, and how flows should be allocated across lanes. These questions are highly relevant for manufacturers, distributors, retailers, wholesalers, consumer goods companies, industrial suppliers, e-commerce networks, and logistics providers operating across multiple products, facilities, geographies, or service regions.

For some businesses, however, footprint and flow design alone is not enough. When production capacity, procurement choices, or inventory positioning materially affect the economics and feasibility of the network, the problem becomes broader. In these settings, network decisions need to be optimized together with additional operating decisions such as what to make where, how much to buy from which suppliers, how much inventory to hold, and when inventory should be built, replenished, or deployed over time. This is especially relevant in multi-echelon manufacturing and distribution networks, seasonal businesses that require prebuild inventory, businesses with constrained plant or distribution-center capacity, and networks where procurement, production, and inventory decisions materially change the best footprint and flow design.

Accordingly, an effective optimization framework should be flexible enough to support both ends of the spectrum: a more basic strategic network design problem when the business mainly needs footprint and flow decisions, and a more integrated network design and operations-planning problem when production, procurement, and inventory must be optimized together with the network itself. That flexibility matters because different industries and business models require different decision scopes.

Once the decision scope is defined, the next question is how these network decisions should be optimized when future operating conditions are not known with certainty. In practice, the economic performance and feasibility of a supply chain network can change materially when demand shifts, transportation rates change, supplier economics move, or available capacity is reduced. This means that the same network design problem can look very different depending on how uncertainty is represented in the model and how response decisions are incorporated into the planning framework.

When dealing with these uncertainties, many supply chain network design and planning efforts still rely on average-case assumptions or scenario-by-scenario comparisons. That approach can be useful for analysis, but it often fails to produce a single executable plan that is optimal before the realized scenario is known. In other words, a network may be designed against an average or representative case, while the actual operating environment later deviates from that assumption in ways that materially affect cost, service, feasibility, or resilience.

The stochastic optimization framework described in this case study addresses that gap by optimizing an Operational Plan and scenario-ready response playbooks together under uncertainty. It is designed to support different decision needs across the spectrum, from a more basic strategic network design problem focused on footprint and flow decisions to a more integrated network design and operations-planning problem that also incorporates procurement, production, and time-phased inventory. In this use case, the framework is applied to network optimization under uncertainty across sourcing, distribution center footprint, transportation flows, procurement, production, and time-phased inventory. Instead of treating volatility as an afterthought, the framework structures uncertainty explicitly and converts it into optimized actions that remain feasible and aligned when conditions shift.

The Challenge

Supply chain networks are exposed to multiple sources of uncertainty at the same time. Demand patterns change by region and season. Transportation rates move unexpectedly. Supplier prices shift, sometimes due to tariff changes. Distribution centers may face reduced capacity or temporary disruption. Suppliers may also become constrained or unavailable.

In real operations, these drivers do not move independently, and a plan that looks optimal under a single forecast or an average case can become expensive or infeasible when reality changes.

Traditional approaches usually fall into one of three categories:

• solving each scenario separately and comparing outcomes
• collapsing uncertainty into a mean or expected-value case
• simulating variability without prescribing coordinated decisions

Each of these can provide insight, but none directly answers the core business question: what is the best network and operating plan to put in place when the future scenario is still unknown, and how should the organization respond when one of the major shifts actually occurs?

The Solution

The stochastic optimization framework addresses this problem through joint optimization of the Operational Plan and the scenario response playbooks.

The process begins by structuring uncertainty across five primary drivers:

• Demand
• Supplier Price
• Sell Price
• Transportation Rate
• Distribution Center Available Capacity

Supplier availability and shortfalls are also treated as part of the broader operating uncertainty affecting the network. These drivers can be enabled or disabled based on business need, and if all are disabled, the case reduces to a traditional deterministic solve.

The framework optimizes decisions as one coherent system across:

• distribution center and supplier selections
• inbound and outbound flows
• sourcing and procurement
• production decisions
• time-phased inventory

The output is not just a recommended network configuration. It includes:

• an Operational Plan to execute before the scenario is known
• scenario-specific response playbooks expressed as operational deltas
• a minimized Risk Response Cost with a driver-level breakdown

This gives the business both a stable base plan and a set of pre-optimized response actions that remain aligned with that plan when tariffs, rates, demand, supplier conditions, or capacity change.

Implementation and Outputs

In a network optimization setting, this framework can be applied to cases such as:

• tariff tightening or easing
• transportation rate spikes and logistics capacity tightness
• seasonal and regional demand variability
• supplier price movement and supply-side risk
• distribution center disruption or reduced capacity
• footprint redesign and network restructuring

For each case, the framework structures the uncertainty, runs the joint solve, and produces outputs that stakeholders can review through both operational and economic views. These outputs include network maps, operational plan views, scenario response views, revenue and cost by scenario, and Risk Response Cost breakdowns.

From an operational perspective, teams receive executable guidance on how sourcing, flow allocation, inventory positioning, procurement, and capacity usage should change under each response scenario while preserving feasibility. From a management perspective, they receive decision-ready tradeoffs across cost, service implications, utilization, and the economics of resilience.

Results and Business Impact

The value of this approach is not only better analysis but better execution. Because the Operational Plan and the response playbooks are optimized together, the business avoids the common problem of generating scenario responses that look attractive on paper but break feasibility somewhere else in the network.

The framework is designed to reduce break-fix replanning, lower operational churn across sourcing and network flows, and support faster response when conditions shift.

It also makes resilience economically explainable. Rather than treating preparedness as a vague premium, the framework minimizes Risk Response Cost and reports its major drivers across procurement, transportation, operations, inventory, revenue, and profit. That allows decision makers to see not only which network choices are more resilient, but also what the incremental cost of that resilience is and where it comes from.

In practical terms, this supports a stronger planning posture:

• more stable execution under volatility
• clearer tradeoffs between cost and resilience
• faster response without improvised replanning
• network decisions tested against uncertainty rather than only average assumptions

Takeaway

Network optimization under uncertainty requires more than testing a few alternative scenarios or solving for an average forecast. The real objective is to make good decisions before the future is known, while ensuring that the organization has aligned, executable responses when conditions change.

This framework addresses that need by combining structured uncertainty, joint optimization, and transparent economics into a single planning approach. By producing an integrated Operational Plan together with scenario-ready response playbooks and a clear view of Risk Response Cost, it helps organizations plan for volatility in a way that is operationally coherent, economically grounded, and ready for execution.

Key Ideas

• Uncertainty is structured explicitly across major drivers rather than treated as isolated what-if cases.
• One joint optimization produces both the Operational Plan and the scenario response playbooks.
• Response actions remain aligned with sourcing, network flows, inventory, and capacity constraints.
• Risk Response Cost is minimized and explained, making the economics of resilience visible.
• Network decisions are evaluated under uncertainty, not only under average assumptions.