Objective:

Develop a robust mechanism for creating a supply chain plan that incorporates variability and addresses the challenges of simulating complex networks. The focus is on ensuring that planning accounts for the inherent volatility of supply chains through AI-driven insights, using a cycle of Plan → Test the Plan → Revise the Plan → Retest until a good plan is established. This approach helps determine whether a plan is effective by simulating the variability in supply chain nodes and continuously improving the plan.

Key Elements:

Traditional Shortcomings:
Current supply chain models (e.g., DDMRP or MRP) often assume a perfect world, failing to account for the real-world variability that supply chains face. These models typically evaluate demand and supply variability separately, resulting in incomplete or unstable plans. This frequently leads to oscillating ordering patterns, creating unstable signals for suppliers and generating inefficiencies across the network.

New Approach to Variability:
The new model simplifies the calculation of variability by moving away from separate demand and supply components. Instead, it introduces a single variability metric based on historical stock on hand vs. expected stock on hand. This unified approach eliminates the complexity of simulating each component separately, providing a clearer view of the overall variability within a supply chain node. By simplifying variability calculation, businesses can achieve better predictability and stability in their plans.

Cycle of Plan → Test → Revise:
This approach introduces an iterative cycle to continuously refine the plan until it meets performance expectations. The process follows these steps:

Plan: Develop an initial supply chain plan based on the best available data and projections.

Test the Plan: Run simulations to test the plan against variability and evaluate its effectiveness. The simulation helps to identify potential issues, such as bottlenecks or unexpected delays, that may arise in real-world execution.

Revise the Plan: Based on the simulation results, revise the plan to address any shortcomings or inefficiencies.

Loop Back to Testing: Retest the revised plan, repeating this cycle until the plan is stable and resilient enough to handle supply chain variability.

This iterative cycle ensures that the plan is not just theoretically sound but practically viable in a dynamic and volatile supply chain environment.

Simulation Strategies:
Traditional models attempt to simulate the entire flow of materials through a network, which can be slow and overly complex. The new approach seeks to overcome this by calculating projections in batches, allowing for faster simulations. Two strategies are considered:

Offset Plan Calculation: Simulate periods for all parts in a batch, calculating offsets between related nodes. For instance, while calculating day 50 for one product, day 70 could be calculated for a product related to the first by a 20-day offset. This approach helps in generating a time-phased plan that takes interdependencies between products into account.

Decoupling Nodes: Completely decouple nodes and simulate each product’s performance independently. This approach does not provide a detailed time-phased result but offers an overall measure of the plan’s effectiveness across a longer period (e.g., a year). This method is particularly useful for understanding broader patterns in network performance without the need for detailed material flow simulations.

AI Technologies:

AI will be crucial in building relationships between nodes and calculating variability. Machine learning algorithms can analyze historical stock data to predict variability patterns, enabling the system to identify where variances are likely to occur. Additionally, reinforcement learning can be applied to test different planning configurations, recommend optimal decisions, and refine the plan based on past performance data. This ensures that the system continuously learns and improves, helping planners make data-driven decisions.

Benefits:

This iterative approach of planning, testing, and revising simplifies supply chain simulations while effectively accounting for variability. It provides more stable and predictable demand signals to suppliers, reducing oscillations and improving overall network performance. By continuously refining the plan through simulations, businesses can proactively evaluate plan effectiveness and resilience, leading to better decision-making, reduced lead times, and a more adaptable supply chain. Ultimately, this approach ensures that businesses are not just creating plans but creating good plans that work in real-world, volatile conditions.

Get Involved

To express your interest in participating in this innovation project, please send an email to: Innovations@b2wise.com quoting “Building a Good Plan”.