Agentic data science with the Data Agent Kit
cloud.google.com·8h ago
AI Research
arXiv CS.AI·Sep 2·5 min read
Discrete-Time MDP Modeling for Multi-Item Capacitated Lot Sizing with Stochastic Demand Timing
TL;DR
A discrete-time Markov decision process (DTMDP) model was developed for multi-item capacitated lot sizing with stochastic demand timing, revealing significant computational challenges. A genetic algorithm (GA) was proposed, achieving an average optimality gap of 3.44% on benchmark instances.
✦ Why It Matters
Consider implementing genetic algorithms for complex lot-sizing problems with uncertain demand timing to improve decision-making efficiency.
Key Takeaways
- Utilize DTMDP for modeling complex lot-sizing problems.
- Implement genetic algorithms for efficient policy search.
- Expect increased computational demands with stochastic timing.
- Use empirical methods to estimate resolution times for unsolvable instances.
How It Works
The model employs a discrete-time Markov decision process (DTMDP) framework, defining states, actions, and costs to manage production and inventory under uncertainty. By focusing on demand-specific decisions, it captures the dynamics of capacity competition and backlog effectively.