Logistics and Supply Chain Optimization with Quantum Computing

  

1. Introduction to Logistics and Supply Chain Management

Understanding logistics and supply chain management is crucial for every business. Supply chains are responsible for delivering products to customers, making them vital to a company’s success. In industries with tight margins, like aerospace or food production, effective supply chain management is even more important. But as companies grow and add new customers, supply chain management becomes more complicated and time-consuming. Logistics and supply chain processes go hand-in-hand. Every step in the supply chain has a logistics component, so logistics and supply chains are interdependent systems. An effective logistics infrastructure is critical for effective supply chain management. This overview focuses on logistics and supply chain management systems, the complexities that go along with them, and the benefits of optimizing them (Phillipson, 2024).

Supply chains and logistics networks are key to a company’s success and growth, and their economic and strategic importance has grown in recent years. A streamlined supply chain can save companies millions of dollars, while deficiencies can drain resources. For example, a company avoided losing $200 million in revenue by rerouting trucks to find missing parts for a new vehicle. Similarly, another company avoided $100 million in losses by meeting demand for a scarce polymer through rapid analysis of the logistics network. As companies grow and their operations become more global, their supply chains can become very large and complex. Rapid changes in supply and demand can wreak havoc on a logistics network. Airlines, for example, had to ground entire fleets when the pandemic wiped out air travel. The complexity of logistics networks often requires a delicate balance between science, mathematics, and creativity to design as optimally as possible. Global competition has heightened interest in logistics and supply chains. This has led to micrologistics innovations in cities, including optimizing package delivery for e-commerce while minimizing traffic disruptions. Technology companies are developing software to help analyze and enhance logistics networks.

1.1. Definition and Importance

Logistics and supply chain management are defined, illustrated, and exemplified in their significance. Logistics management is defined as the planning, implementation, and control of the efficient movement and storage of goods, services, and related information from point of origin to point of consumption. All activities that make up logistics management can be divided into five components: transport, warehousing, inventory management, order fulfilment, and logistics system design. Logistics management has a pivotal role in the economy, coordinating the physical flow of goods from suppliers to customers through different modes of transport and integrating various logistics activities (Phillipson, 2024). Supply chain management is defined as the planning, implementation, and control of supply chain activities to achieve the fixed supply chain goals of high efficiency and low operating costs. All activities that make up the supply chain management can be divided into five components: supply chain network design, procurement, production planning, transportation planning, inventory planning, and sales planning. The goal of each supply chain is to minimise the total cost of the supply chain while meeting the demand of each customer. A supply chain is a network of upstream suppliers, manufacturers, distributors, and downstream retailers and customers. Effective logistics and supply chain management make a significant contribution to a company’s profitability, ensuring that goods are moved in the most cost-effective way while still meeting customer service requirements. Real-world examples illustrate that logistics and supply chain management cannot be ignored even by the largest players. This illustrates the vital nature of these two domains, alongside essential definitions and clarifications. As such, it would seem that an awareness of contemporary trends in logistics and supply chain management optimisation is required before the space is extensively explored. Traditionally, logistics and supply chain management are considered in matrix technology-dominated paradigms, where material flows in the supply chain, internal logistics, and at individual nodes of the network are actively researched. This forms the groundwork for the investigation of the quantum computing approach and its applications within logistics and supply chain management, beginning with a critical overview of the traditional optimisation approaches and their limitations. Focussing on quantum computing, currently perceived as the most promising next computing paradigm after classical von Neumann computing, is justified. The space of computationally intensive optimisation problems, including logistics and supply chain management issues, most suitable for hybrid quantum computing solutions is identified. By framing the concepts of logistics and supply chains, the key questions and answers relevant to the optimisation domain are outlined.

2. Challenges in Traditional Optimization Methods

In logistics and supply chain management, the quantitative optimisation of a network of interconnected processes and resources is central to maximising operational efficiency, minimising costs, and maintaining agreed service levels. Consequently, this field has been of growing interest to both academia and industry. Various challenges associated with logistics and supply chain optimisation, however, complicate decision-making processes (Phillipson, 2024). On an operational level, route planning and vehicle scheduling problems seek to minimize transportation costs while meeting demand, adhering to time-windows and vehicle constraints, and avoiding inefficient driving distances. These problems are typically NP-hard, and infeasible solutions can result in lengthy vehicle routes that exceed legal driving-hour limits or unmet demand. Furthermore, large fleets and use of subcontractors can lead to the emergence of complex network structures, where a single vehicle is assigned to multiple stops at different locations, creating a bottleneck that necessitates the need for a vehicle to visit multiple sites in sequence.

As a supply chain grows more complex, it is observed that the problems involved become larger and more difficult to solve. In addition to the number of relevant variables and process steps, the nature of the dynamics plays a crucial role. Modern supply chains are likely to rely on a temporary alliance among several companies, and each participant must constantly negotiate a balance between various objectives, such as minimising cost versus lead time, and robustness versus efficiency. The outcome of these negotiations in turn affects all decisions that have been made upstream, necessitating an ongoing reconsideration of previously fixed parameters. Decisions made on one level can therefore lead to suboptimal solutions on another level, and the resultant accumulation of errors increases operational costs. A similar effect arises when attempting to contain the system's complexity by isolating subsystems. Local optimisations tend to be satisfactory only for tightly controlled test scenarios, which rarely occur in practice. The dynamically evolving environment of an optimisation problem leads to a gradual degradation of its quality.

Numerous optimisation tasks are time-critical, meaning that the quality of a solution must be continuously monitored, and corrective actions taken, while new data continually flow into the system. Changes in the environmental conditions, such as customer requests or vehicle breakdowns, can occur at any moment and must be processed in real-time. Many traditional optimisation strategies rely on a discretisation of time, meaning that solutions are only adapted at fixed intervals. As a result, poor performance is observed in reconciling transportation plans with real-time data, not to mention incorporating these data into a wider network of planning. Other strategies fix most parameters for the duration of an optimisation run, leading to the need for alternative solutions when data outside the predetermined range are encountered, for example, in the event of unforeseen temporary changes in capacity or demand.

Adaptational strategies that can accommodate ongoing changes in the market, for example, by automatically updating the system's expectations regarding the frequency and size of changes, are an emerging area of research. However, a fully adaptable system remains elusive. Even slight modifications can render control mechanisms ineffective, reflecting the pathological behaviour of chaotic systems. As the quality of a solution degrades, it is impossible to ascertain how its reconciling might affect other decisions. As a rule, feedback loops between different subsystems are neglected to avoid this binding complexity.

2.1. Complexity and Scale

Logistics and supply chains are vital to business operations, yet optimally designing and managing them is complex. As the scale of logistics operations increases, so does the complexity of achieving efficiency. These complexities stem from the interplay of many variables, some of which are difficult to quantify. Further complications arise because most variables may only change in discrete steps. Additionally, many factors have an uncertain nature, including costs and demand, which complicates scalability. Further complexity is added by continually changing market conditions. Many traditional models assume static environments and fail to adequately represent the dynamic nature of the industry. While linear models can tackle many static problems, numerous non-linear considerations render them ineffective (Phillipson, 2024).

Heuristic approaches like genetic algorithms and simulated annealing can yield satisfactory solutions but may be time-consuming or converge on local minima, failing to find the best option. The growth of global supply chains has made management strategies more complex by adding layers of complexity. Decisions optimal for a local operation may be detrimental on a larger scale. For example, a facility closing may reduce costs locally but increase transportation costs elsewhere. A global perspective is required to assess the interdependence of subsystems. Growing complexity, uncertainty, and the interplay of continuous and discrete variables mean that even large corporations with sophisticated models often rely on basic approaches to route planning and scheduling.

Logistics and supply chain stability depend on complex networks. Disruptions can ripple through the network, and the failure of one component may lead to the failure of others. Containing disruptions is thus challenging. Many complex networks display a characteristic structure with various properties, including the presence of hub nodes. The loss of a hub may disconnect the network, and preventing a cascade of failures is difficult. Multi-tier networks often develop, resulting in complicated interactions between subsystems. In this increasingly intricate environment, sophisticated quantitate optimization techniques are necessary but currently lacking.

3. Quantum Computing Fundamentals

Quantum mechanics is the foundation of quantum technologies. Proponents of quantum computers promise they can offer informational processing advantages over the best conventional computers for certain problems. Complex systems in nature, for instance, are difficult to model with classical computers, whereas quantum states are naturally given by the same mathematics as quantum mechanics. Therefore, quantum computers would be ideal for simulating other quantum systems. Many relevant systems span different disciplines, from chemistry to high-energy physics, from condensed matter to material sciences, and from nanotechnology to biophysics. Outside of science, the financial world deals with quantum Monte Carlo methods and path integrals. Other problems like weather prediction rely on modelling complex classical systems with high-dimensional phase space. All such problems are NP-hard for classical computers and would benefit from possible speedup with quantum computers (Phillipson, 2024).

A quantum bit, or qubit, is the fundamental unit of information in a quantum system. A qubit can be in a state |0⟩, |1⟩, or in a quantum superposition of both states, which can be represented by a state vector |ψ⟩ in Hilbert space. The pure state of a qubit can be represented on the Bloch sphere. A perfectly mixed qubit state is described by a density matrix; a completely mixed state corresponds to a vector pointing to the center of the Bloch sphere. A 2N-dimensional Hilbert space encodes all quantum states of N qubits. A classical bit can be either 0 or 1, while a qubit can be 0, 1 or in any combination of the two states, reflecting the parallelism in quantum computing. Many qubit systems exhibit quantum gates, which are peculiar operations acting on qubits and can modify the quantum state. Quantum gates have a representation in terms of unitary 2×2 matrices acting on the qubit state vector (|ψ⟩). For instance, the Hadamard gate creates a superposition of the state |0⟩. A combination of quantum gates creates a quantum circuit that can perform an algorithm on qubits. Quantum entanglement can arise from a joint operation on several qubits. An entangled state cannot be separated into single-qubit states and can lead to a Bell inequality violation. The EPR experiment proposed by Einstein, Podolsky, and Rosen illustrates the non-locality nature of entangled states: measuring one qubit determines the state of the other with certainty, regardless of the distance separating them.

3.1. Quantum Superposition and Entanglement

To understand the principles of quantum computing it is pivotal to comprehend quantum superposition and quantum entanglement. Quantum superposition describes an intrinsic behaviour of quantum systems, which allows for quantum bits (qubits) to exist in multiple states simultaneously. Although standard bits in classical computing can either be in “0” or “1” states, qubits can be in “0”, “1” or in a linear combination of both states. This characteristic of qubits allows quantum computers to vastly increase the processing power and as a result execution speed of numerous applications. Quantum superposition is at the base of the exponential growth of quantum states in a many-qubit system. A system of n qubits can exist in 2^n different states, therefore computations that classical computers struggle with can be realised efficiently on quantum systems.

Another significant quantum phenomenon is quantum entanglement, which explains how qubits can be interconnected in such a way that the measurement of one qubit affects the measurement of the second, even if they are separated in space. The degree of entanglement determines how much the individual quantum states of the qubits are altered by local transformations. Quantum gates exploit this property to enhance the computational efficiency and speed of the performed operations. In the context of quantum computing, quantum systems and algorithms designed to operate on those systems use superposition and entanglement to achieve speed-up compared to their classical counterparts.

Many logistical and supply-chain optimisation problems are complex by nature, thus it might not be feasible to find an optimal solution using classical processors. Quantum processors are expected to be able to use their quantum characteristics to find optimal solutions for such problems in a reasonable execution time. Problems, such as the well-known Travelling Salesman Problem (TSP) or its variations, where a set of cities and corresponding distances between them need to be solved, are NP-hard for classical computation (Phillipson, 2024). There are, however, many other applications in operations management that can benefit from quantum computing.

4. Applications of Quantum Computing in Logistics and Supply Chain Optimization

Keeping goods flowing across the globe poses an ever-growing set of logistical challenges for companies as the commodity routing, resizing, and scheduling optimisation problems get more complex with larger, faster-moving and more varied supply chains. These challenges were massively exacerbated by the pandemic in 2020. They are being pursued most rigorously by the largest players, for whom even small improvements can yield savings in the billions of dollars. At the same time, the rising global demand for logistical services strains often already overburdened transport networks and a scarcity of resources, requiring a more efficient use of transport resources to keep the systems running smoothly.

Quantum computing promises a paradigm shift in computational capabilities by leveraging quantum phenomena such as superpositioning and entanglement. Consequently, quantum computing is expected to outperform classical computing in solving specific problems, including combinatorial optimisation problems that are ubiquitous across entire sectors, including transport, logistics, traffic, scheduling and supply chains. Quantum technologies and, in particular, quantum algorithms have been demonstrated to improve classical approaches and KPIs in real-world logistics and supply chain scenarios, including quantum-enhanced route optimisation for goods deliveries, waste collection in public transport and large-scale resource allocation with quantum computers on a chip.

Logistics and supply chains are quintessential complex systems composed of many interdependent parts, which are the basis of contemporary society and global economy. This complexity renders them fragile as disruptive events, such as vehicle failures, unfulfilled deliveries, accidents, system malfunctions, telecommunication breakdowns or pandemics, can have dramatic cascading effects, leading to the breakdown of entire subsystems. Thus, a tight balancing of efficiency and resilience is required for such systems. A goal of the ongoing quantum developments and a pathway for collaborative efforts between academia and industries is to exploit the quantum capabilities for innovative solutions to critical challenges in logistics and supply chains.

4.1. Route Optimization

4.1. Route Optimization

Few applications can have such a transformative effect as logistics and supply chain optimization with new quantum computing technology. Traditional methods for route planning and optimization work well in straightforward applications, such as organizing the delivery of a parcel to a single recipient. However, they quickly lose efficiency when faced with the complex network of many pick-up points, drop-off points, delivery constraints, and vehicle capabilities. Digitally simulated quantum computing will excel in these situations, achieving results impossible with classical systems due to the exponential growth of solution complexity (Phillipson, 2024).

Quantum algorithms have been shown to outperform their classical counterparts in identifying routes that minimize travel time and costs by taking advantage of quantum superposition, where a network of quantum processors can analyze multiple viable routes simultaneously. Delivery services that harness the power of this new computing technology can expect a dramatic competitive advantage. Fleet management can be optimized with similar quantum-enhanced algorithms, speeding up computation time for vehicle routes and schedules despite varying service times, multiple job pickups or drop-offs, and hard time windows. Implementing quantum route planning would improve overall supply chain efficiency by reducing costs and energy consumption while allowing for rapid adaptation to unforeseen events such as traffic disruptions or vehicle breakdowns.

Fortunately, practical implementation is closer than it may seem. In 2021, Dijkstra’s algorithm for the shortest path in a graph was successfully tested on a quantum processor using an approach that blends quantum and classical computing. In July 2023, new algorithms were presented that plan fleet vehicle routes for many jobs, taking advantage of recent discoveries in combinatorial QAOA circuit design. In numerical simulations, they find plans up to 34 percent better than comparable classical methods, demonstrating substantial improvements for more significant planning problems. With such quantum-enhanced route planning, businesses could realize practical benefits, as evidenced by the industrial implementation of Classical Fleet Vehicle Planning Quantum applications.

5. Compelling Case Studies and Inspiring Success Stories

This paper presents a collection of case studies and success stories that demonstrate the effective application of quantum computing in logistics and supply chain optimization. Each case study highlights a different scenario and context in which quantum technologies have been employed. The outcomes of these implementations show significant improvements in supply chain and logistics KPIs, such as efficiency, costs, and decision-making. These examples provide credible proof of the real-world impacts of quantum computing and highlight the value of utilizing these technologies. The aim is to encourage and inspire further exploration and investment in quantum strategies across a broad spectrum of industries. Specific metrics and data are provided from these implementations, showcasing the tangible benefits achieved. By summarizing these quotes and examples, the intention is to illustrate how the potential of quantum computing can be harnessed widely. This discussion concludes by illustrating a pathway forward for practitioners and researchers alike.

Taking strides towards a greener and more efficient supply chain, logistics companies face complex challenges such as fuel price fluctuations, climate regulations, evolving customer demands, and fierce competition. Amidst these trials, the pressure for smarter, and thus more complex, decision-making processes in the supply chain increases. From a technical point of view, these decisions often boil down to intricate optimization problems that can involve thousands or even millions of variables and constraints (Phillipson, 2024). Classical algorithms struggle to solve these problems within a reasonable timespan or might fail altogether, especially when real-life problems involve multiple conflicting objectives—such as cost reduction while increasing service level and decreasing CO2 emissions. This is when quantum computing comes into play as a promising new paradigm that could tackle such hard optimization problems faster and more efficiently than classical computers do.

References:

Phillipson, F. (2024). Quantum Computing in Logistics and Supply Chain Management an Overview. [PDF]