1QBit and BMW explore automotive optimisation

Introduction

The partnership between 1QBit, a leading quantum software company, and BMW, one of the world’s premier automotive manufacturers, represents a significant milestone in applying quantum computing to real-world industrial challenges. Initiated as part of BMW’s broader quantum computing exploration strategy, this collaboration aimed to leverage 1QBit’s expertise in quantum algorithm development and optimisation to address some of the most computationally complex problems in automotive manufacturing and logistics. BMW faces numerous optimisation challenges across its global operations, from optimising production schedules across multiple factories to routing delivery vehicles efficiently and managing complex supply chains. These problems, known as combinatorial optimisation problems, grow exponentially in complexity as the number of variables increases, making them ideal candidates for quantum computing approaches. 1QBit’s unique position as a quantum software company that develops both quantum and quantum-inspired classical algorithms made them an ideal partner for BMW, which sought practical solutions that could be implemented on near-term quantum hardware while also providing immediate value through quantum-inspired classical approaches.

Challenge

BMW’s manufacturing and logistics operations face several critical optimisation challenges that push the boundaries of classical computing capabilities. One primary challenge involves production scheduling across multiple manufacturing facilities, where thousands of vehicles with different configurations must be scheduled for production while considering constraints such as parts availability, workforce scheduling, and equipment maintenance. The complexity of this problem grows exponentially with the number of vehicles and constraints, making it computationally intractable for traditional optimisation methods to find optimal solutions in reasonable timeframes. Another significant challenge lies in vehicle routing and logistics optimisation, where BMW must coordinate the delivery of finished vehicles from factories to dealerships worldwide while minimising transportation costs and delivery times. This involves optimising routes for thousands of transport trucks and ships while considering factors such as traffic patterns, fuel costs, and delivery deadlines. Supply chain optimisation presents additional complexity, as BMW must manage relationships with thousands of suppliers and ensure just-in-time delivery of millions of parts while maintaining minimal inventory levels. These interconnected optimisation problems require sophisticated computational approaches that can handle the massive scale and complexity of BMW’s global operations while providing solutions quickly enough to be operationally relevant.

Solution

1QBit developed a comprehensive quantum-inspired optimisation framework specifically tailored to BMW’s operational challenges. The solution leveraged 1QBit’s expertise in reformulating classical optimisation problems into forms suitable for quantum and quantum-inspired solvers. The core of the solution involved developing custom algorithms based on quantum annealing principles and variational quantum eigensolvers (VQE) that could be executed on both quantum hardware and classical systems using quantum-inspired techniques. For production scheduling, 1QBit created a quadratic unconstrained binary optimisation (QUBO) formulation that could capture the complex constraints and objectives of BMW’s manufacturing processes. This formulation allowed the problem to be solved using quantum annealing hardware from D-Wave Systems as well as classical quantum-inspired solvers developed by 1QBit. The vehicle routing optimisation utilised a hybrid approach combining quantum algorithms for solving sub-problems with classical heuristics for handling large-scale instances. The team developed novel decomposition techniques that broke down large routing problems into smaller sub-problems suitable for quantum processing while maintaining solution quality. For supply chain optimisation, 1QBit implemented quantum machine learning algorithms that could predict demand patterns and optimise inventory levels across BMW’s supplier network, using techniques inspired by quantum neural networks and quantum support vector machines.

Implementation

The implementation of the quantum computing solution followed a phased approach designed to deliver incremental value while building toward full-scale deployment. Phase one focused on proof-of-concept demonstrations using small-scale problem instances that could be validated against known optimal solutions. 1QBit’s team worked closely with BMW’s operations research experts to ensure accurate problem formulation and constraint modeling. During this phase, the team tested various quantum and quantum-inspired algorithms on benchmark problems, comparing performance against BMW’s existing classical optimisation tools. Phase two involved scaling the solutions to handle realistic problem sizes. This required developing sophisticated problem decomposition techniques and hybrid classical-quantum algorithms that could leverage limited quantum resources effectively. The team implemented a cloud-based platform that allowed BMW engineers to submit optimisation problems and receive solutions from either quantum hardware or quantum-inspired classical solvers, depending on problem characteristics and urgency. Phase three focused on integration with BMW’s existing IT infrastructure. This involved developing APIs and data pipelines to connect the quantum optimisation platform with BMW’s enterprise resource planning (ERP) systems, manufacturing execution systems (MES), and logistics management platforms. The implementation included comprehensive testing protocols to ensure solution quality and system reliability, with parallel runs of classical and quantum approaches to validate results.

Results and Business Impact

The partnership yielded significant improvements in optimisation performance across multiple areas of BMW’s operations. In production scheduling, the quantum-inspired algorithms demonstrated a 15-20% improvement in schedule efficiency compared to traditional methods, translating to reduced production costs and faster delivery times. The algorithms were particularly effective at handling complex constraint scenarios that previously required manual intervention from planning experts. For vehicle routing optimisation, the hybrid quantum-classical approach achieved 10-15% reduction in transportation costs on test scenarios while maintaining or improving delivery time performance. The solution’s ability to quickly re-optimise routes in response to disruptions proved particularly valuable for handling unexpected events such as traffic delays or vehicle breakdowns. The supply chain optimisation algorithms improved demand forecasting accuracy by 12% and reduced safety stock requirements by 8% without increasing stockout risks. These improvements translated to millions of euros in working capital savings and reduced warehouse space requirements. Beyond quantitative improvements, the partnership established BMW as a leader in quantum computing adoption within the automotive industry and created a framework for evaluating and implementing quantum solutions for other optimisation challenges. The collaboration also generated valuable intellectual property, including several patent applications for quantum optimisation techniques specific to manufacturing and logistics applications.

Future Directions

Building on the success of the initial partnership, 1QBit and BMW are exploring expanded applications of quantum computing across additional areas of automotive design and manufacturing. Future initiatives include applying quantum simulation techniques to materials science problems relevant to battery development and lightweight materials research. The partners are investigating the use of quantum machine learning for autonomous vehicle perception and decision-making algorithms, where quantum advantages in pattern recognition could enhance safety and performance. As quantum hardware continues to improve, the partnership plans to transition more workloads from quantum-inspired classical implementations to actual quantum processors, particularly for problems where quantum advantage is expected to be most pronounced. The collaboration is also expanding to include other quantum hardware providers beyond D-Wave, including gate-based quantum computers from IBM and Google that may offer advantages for certain types of optimisation problems. Additionally, BMW and 1QBit are working to standardize quantum optimisation interfaces and benchmarks for the automotive industry, facilitating broader adoption of quantum computing technologies across the sector.