Our Offerings: A Strategic AI Agent with a Brain
THE TOOL
THE MOTIVATION
THE GOAL
To use an algorithm able to weigh countless complex interactions and dynamics to recommend the best action
Deep Reinforcement Learning
AI agent-optimized electrification
Fractal harnessed the capabilities of the AnyLogic simulation platform’s transition simulator. We used this simulation to train a BRAIN (Basic Recurrent Artificial Intelligence Network), a Deep Reinforcement Learning (DRL) AI agent. As opposed to a predictive model, a BRAIN can react to a dynamic and novel environment in real time, recommending the best action at any given moment. Scaling things up to run hundreds of simulations in parallel, we expedited our AI agent’s development progress with a viable result in only a few weeks. Our client’s subject matter experts then rigorously reviewed and approved the final electrification plan for the fleet of dump trucks. We then leveraged the brain to explore goals optimized to align with various possible futures or different objectives, ensuring no matter what happens, our client was set up for an electrifying and brilliantly optimized future.
To perform training with hundreds of parallel simulations testing nigh endless possibilities
Enhanced reliability and recommendation robustness
Simulation Training
Increased credibility and confidence in plan feasibility
To ensure accuracy and validate the electrification plan
Subject Matter Experts
To provide a trained AI agent from simulated scenarios that can be used in an ongoing manner as key dynamics change over the next 10+ years
Data-driven decisions for fleet electrification now and in the future
Custom AI “BRAIN” Solution
To achieve a balance between eco-responsibility and transition costs
Reduced emissions and achievement of ESG target
Environmental & Cost Goals
© 2023 Fractal Analytics Inc. All rights reserved
5
Made with FlippingBook - PDF hosting