Master Thesis Student

Stochastic MPC and Reinforcement
Learning for Robust Electric Truck
Mission Planning Under Uncertainty

About us
Sustainability, including climate change, is one of the key challenges of our generation.
Our contribution is to offer leading transport and infrastructure solutions that enable
societies to prosper in a sustainable way. At Volvo Group, we are committed to the
ambitions and climate goals of the Paris Agreement. From a lifecycle perspective, most
emissions occur during the use phase of our products. Therefore, our priority is to develop
solutions that reduce carbon emissions from transportation. Today, Volvo is the market
leader in offering a full range of electric trucks.
We are the Transport Productivity team under Vehicle Efficiency and Productivity
at Group Trucks Technology (GTT). Our work focuses on optimal charge planning for
Battery Electric Trucks (BEVs), where the goal is to minimize downtime and maximize
operational uptime. Using predicted route information such as topography, road curvature,
and speed limits, we intelligently suggest charging stations along the route to ensure
efficient energy use and seamless trip execution.

Thesis Background
For electric trucks to match or exceed the utility of conventional fossil-fuel vehicles, intelligent
trip planning is necessary. One of the key limitations of electric vehicles (EVs)
is the comparatively long charging time of lithium-ion batteries, which reduces vehicle
uptime and extends mission duration. Furthermore, the limited availability of charging
stations further contributes to range anxiety, where drivers fear running out of battery
before reaching a station. Effective trip planning can significantly mitigate this issue.

Motivation and Problem Statement
The objective of this thesis project is to design and evaluate a robust stochastic optimization
scheme for long-range electric truck operations. The primary goal is to minimize
the total mission duration (travel + charging + driver rest time) by finding an optimal
sequence of charging locations and driver rest stops while addressing real-world uncertainties.
The ongoing research at Volvo, including a previous master thesis [1], has explored
this problem using techniques such as hierarchical optimization and nonlinear integer
programming (NLIP). A remaining challenge is to incorporate driver rest periods and
real-world uncertainties into optimization to improve planning robustness.
In this project, charge planning will be formulated as a full-horizon Dynamic Programming
(DP) problem, where the objective is to find an optimal sequence of charging over
the entire mission. However, solving DP exactly becomes computationally intractable
when uncertainties such as traffic, queue times, or weather conditions are included in the
model. To address this, Reinforcement Learning (RL) will be used as an Approximate
Dynamic Programming (ADP) method. RL will be trained through simulations to estimate
the long-term value of different charging and rest strategies, effectively learning a
policy that generalizes to new conditions. This combination of DP-RL enables optimal
or near-optimal planning over large horizons while remaining computationally efficient.
Objective or Research Question
• What are the limitations/challenges and quantifiable benefits of such a solution?
• How to formulate the optimization problem? Reasoning for simplifying or excluding
certain dynamics/parameters.

Deliverables
• Formulating the trip planning problem within an optimal control framework such
as Model Predictive Control (MPC) and Reinforcement Learning (RL).
• The influence of uncertainty models such as traffic variability on mission planning
robustness will be systematically analyzed.
• The performance of a hybrid MPC–RL framework will be benchmarked against a
standalone MPC/heuristics to determine improvements in mission efficiency and
disturbance resilience.

Suitable Background
• Interest in optimal control
• Master students in Automotive, Control, Mechatronics, Engineering Physics or similar
• Good skills in scripting languages: e.g., Matlab or Python

Thesis Level
Master

Language
English

Starting Date
Spring 2026

Physical Location
Mostly at Volvo Lundby

Supervisor
Academic Supervisor:
Nikolce Murgovski (nikolce.murgovski@chalmers.se)
Mohamed Abrash (abrash@chalmers.se)
Volvo Contact Persons:
Olof Lindgärde (olof.lindgarde@volvo.com)
Fatemeh Mohammadi (Fatemeh.mohammadi@volvo.com)

References
[1] Isac Borghed. Hierarchical optimization for charge planning and thermal management
of battery electric trucks. 2024.