Real-time adjustments
Continuous quality measurement
Seamless quality assurance
Optimized efficiency
The challenge
Optimizing Cheetos production with AI
Ensuring consistent quality in Cheetos production is complex due to variations in raw materials, equipment behavior, and environmental factors. PepsiCo needed a solution to maintain precision, optimize efficiency, and minimize waste across its manufacturing lines.
Key challenges
Raw materials, environment, and equipment vary
Equipment changes over time, affecting output
Machines of the same model behave uniquely
Ensure precision while reducing waste
The solution
AI-powered production optimization
AI-powered optimization
AI agent optimizes yield
PepsiCo and Microsoft AI partnership
Used Deep Reinforcement Learning (DRL)
Smart simulation and training
Deep Neural Network simulator
Optimized via Machine Teaching
AI trained on real data
Implementation approach
1
AI training and setup
Developed process simulator
Refined AI through iterative testing
Used real data for model accuracy
2
Reward function tuning
Applied machine teaching logic
Adjusted parameters for precision
Aligned AI with production goals
3
Real-world deployment
Integrated AI into production
Enabled real-time adjustments
Enhanced efficiency and consistency
The impact
Maximising quality and efficiency with AI
Real-time adjustments
AI fine-tunes extruder settings
Ensures consistent quality
Reduces variations
Seamless monitoring
Continuous quality checks
No production disruptions
Minimal manual intervention
Optimized performance
Higher efficiency, less waste
AI-driven yield improvement
Reliable, top-quality output
Looking ahead
Scaling AI adoption
Expand AI integration across production lines
Continuous improvement
Enhance algorithms for greater precision
Sustainable efficiency
Drive long-term optimization and waste reduction
