Smarter Cold Storage: Delivering 20.4% Energy Reduction with AI
NewCold is a global company specialising in automated cold chain logistics. Their long-term mission is to develop world-class, energy-efficient warehouses.
To help them achieve their goals, Qubiz worked on an innovative AI solution based on reinforcement learning.
Services
IT Consulting
IT Consulting
AI Development
AI Development
Data Science
Data Science
Collaboration Impact
20,4% Energy Efficiency
Decrease in average price consumed/kWh compared to baseline
Scalability Across 15 Warehouses
Scaling this solution for warehouses around the world
Savings of EUR 47,000
Estimated yearly savings for a single warehouse
Our challenge
Energy costs are rising all around the world. Monitoring energy and cost usage across an entire warehouse is time-consuming. NewCold wanted to explore AI solutions to do this more efficiently.
Our approach
As their long-term strategic partner, we developed an intelligent agent based on reinforcement learning.
When released in a real environment, the system is ready to work with live data (such as temperature and humidity) to minimise energy costs and waste generated by refrigerator cooling systems, picking locations, and more.
For this purpose, we used a deep learning model, LSTM, used for forecasting based on time series data. In this way, we simulated warehouse operations, which the agent uses as a training environment, learning optimal policies for energy cost reduction.
AI Agent Development
AI Agent Development
The Qubiz R&D team provided AI consultancy and implementation
Exploration Phase
Exploration Phase
Using energy data collected from the warehouse, the agent initially goes through a trial-and-error phase, learning based on feedback
Exploitation Phase
Exploitation Phase
Once the agent has learned the optimal energy policies, it makes autonomous decisions regarding how energy is consumed, based on the current state of the cooling system
Technologies
Business outcomes
The agent makes autonomous decisions regarding energy consumption based on the current state of the cooling system, resulting in major cost savings.
The AI agent is able to solve similar problems for any particular warehouse, even if it’s located in a different area or is brand new
Going from a manual process to an agent capable of working independently to boost sustainability and save around EUR 47,000/year for a single location
The AI agent functions accurately and is trained to work well from day 1 in a new environment
As their strategic partner, we help NewCold fulfill its mission of being the most innovative and sustainable cold storage company in the world
