Problem:
Long-term manufacturing investments were being made on a single forecast in an unpredictable world. When demand shifted or assets stalled, decisions became fragile and risky.
Solution:
AstraZeneca and Decision Lab built a simulation platform to support strategic capacity planning. It allowed teams to test decisions across thousands of scenarios and move toward an antifragile supply chain.
Results:
- Reduced reliance on fragile spreadsheets and enabled faster scenario testing.
- Improved visibility of trade-offs and uncertainty.
- Strengthened alignment between R&D and manufacturing.
- Enabled confident, data-backed investment decisions.
- Shifted planning from reacting to volatility to using it as an advantage.
Introduction: planning under uncertainty
Planning manufacturing capacity in pharma is trying to prepare for a future you cannot fully see.
Years before medicine reaches patients, investments must be approved, facilities expanded, and production prepared. But at that point, clinical outcomes are still uncertain, demand can shift, and timelines often change.
Uncertainty is not unique to pharma. Across industries, supply chains face growing disruption: 9 out of 10 leaders consider it the number one risk, and 63% of supply chains are classified as fragile because they lose value under stress.
For AstraZeneca, a global biopharmaceutical company, this raised a difficult question: how do you commit to major manufacturing investments when so much is still unknown? The answer required rethinking strategic capacity planning to better reflect real-world volatility.
Rather than relying on a single forecast, the company chose a different path. Together with Decision Lab, AstraZeneca decided to build a strategic simulation platform powered by AnyLogic to explore multiple possible futures and engineer greater antifragility into its supply chain.
Problem: limitations in strategic capacity planning
Like many large organizations, AstraZeneca relied on complex spreadsheet-based workflows for long-range planning. Over time, this approach became increasingly fragile and limited the effectiveness of its strategic capacity planning.
Planning decisions relied on a single deterministic forecast in a world that rarely follows a linear path. When demand shifts, portfolios change, or supply chains face disruption, a single forecast quickly loses relevance.
This created several issues:
- Slow scenario testing and heavy manual effort.
- High risk of version control problems and copy-paste errors.
- Limited ability to understand uncertainty or stress-test strategies.
- Difficulty aligning Research and Development (R&D) portfolio decisions with manufacturing capacity investments.
In pharma, R&D refers to the process of discovering, testing, and developing new medicines before they reach patients.
Overall, the problem was not a lack of data or expertise. It was unpredictability itself. Assets stalled, demand shifted, and disruptions hit without warning.
A single forecast could not capture that volatility. The business needed a way to test decisions across multiple possible futures and move toward a more antifragile supply chain.
The goal was not only to withstand disruption but also to turn volatility into a strategic advantage by using uncertainty itself as a basis for better decisions.
Solution: a modular simulation approach
Together, Decision Lab and AstraZeneca designed a strategic simulation platform built in AnyLogic to transform strategic capacity planning.
At its core, the solution connected three distinct but interdependent models:
- Portfolio model Simulated the lifecycle of R&D assets as they progress through development stages.
- Demand model Translated portfolio outputs into quarterly forecasts of required active ingredients.
- Capacity model Evaluated manufacturing scenarios and investment strategies under varying conditions.
Each model ran independently in AnyLogic Cloud and was orchestrated through a custom web interface built with React and Node.js. This modular architecture allowed each model to evolve without breaking the overall system.
Moving from single-point to range-based planning
One of the most important design decisions was shifting from a single forecast to range-based planning.
Using Monte Carlo experiments in AnyLogic, the team ran thousands of simulations with varied inputs. Instead of asking, “What is the forecast?” planners began asking what range of possible futures could emerge and which strategies would perform best across that range.
The platform helped them visualize:
- Multi-run distributions and heat maps.
- Scenario branching trees.
- Side-by-side KPI comparisons.
Users could filter simulation runs and pass selected outcomes forward to the next model in the chain. This allowed AstraZeneca to test capacity investments across multiple plausible futures and strengthen its antifragile supply chain.
Why AnyLogic
AnyLogic was chosen because the project required a system built for uncertainty, not just a better forecast. It supported this in two keyways:
- The ability to run Monte Carlo experiments at scale. This allowed the team to move beyond a single forecast and strengthen strategic capacity planning across a full range of possible outcomes.
- A cloud-based modular architecture. Portfolio, demand, and capacity models could operate as independent services while still functioning as one connected system.
This combination made it possible to build a scalable and explainable system rather than a single monolithic model.
What they built was more than a simulation tool. It was a practical decision system that linked what happens in the R&D pipeline to real manufacturing investments and supported a more antifragile supply chain. This gave leaders a clearer understanding of how the whole system behaves before committing to long-term decisions.
Read also: How GSK used simulation to design and optimize a new biopharmaceutical manufacturing facility, reducing capital costs and improving production planning.
Results: progress toward an antifragile supply chain
The impact of the platform went beyond faster simulations. It fundamentally changed how planning conversations happen and strengthened strategic capacity planning across the organization.
Automated workflow
Manual spreadsheet handling was replaced with an end-to-end simulation pipeline.
- Reduced risk of data errors and fragile links.
- Freed internal experts to focus on strategic analysis instead of data consolidation.
A shared decision language
The visual interface and scenario comparison capabilities brought real transparency into the planning process.
Instead of different teams working from separate spreadsheets or interpretations, both technical experts and leadership could look at the same outputs and the same assumptions. Trade-offs were visible. Uncertainty was clear. Important decisions were no longer buried in complex files but discussed openly with a shared understanding of the numbers behind them.
Data-driven confidence
By planning within a range rather than around a single number, leadership could:
- Stress-test high-value investment decisions.
- Understand downside and upside scenarios.
- De-risk long-term technology and capacity commitments.
Most importantly, the platform supported a shift from reactive planning to proactive strategy and advanced the development of a more antifragile supply chain. Instead of simply trying to survive disruption, AstraZeneca could evaluate how different decisions perform under volatility.
The shift was not just technical. It represented a change in mindset. Planning is no longer about predicting a perfect future. It is about preparing for many possible ones.
The case study was developed and presented by Disa Ray from Decision Lab and Gareth Alford from AstraZeneca at the AnyLogic Conference 2025.
The slides are available as a PDF.