Application Of Quantum Annealing in Factory Layout Planning

Abstract

Quantum computing offers a revolutionary approach to computation by leveraging quantum bits or qubits. Quantum computers can perform certain calculations exponentially faster than classical computers. In the context of factory layout planning, quantum annealing, a specific quantum computing approach, can efficiently optimize complex problems by exploring multiple solutions simultaneously and quickly finding the optimal configuration for factory layouts. By harnessing quantum annealing’s ability to navigate vast solution spaces rapidly, factory planners can enhance efficiency, reduce costs, and improve overall productivity in designing optimal factory layouts. This thesis aims to prove that not only factory layout planning is achievable using quantum annealing, but it is also many folds faster than classical Monte-Carlo simulation techniques. This thesis proves that quantum annealing can be used to generate multiple optimal or near-optimal layouts which can help in the early stages of factory layout planning. By utilizing D’Wave’s QUBO formulation, a factory can be divided into a number of positions and, along with the functional units, can be described as a graph, with nodes mapping to qubits and the edges mapping to couplers between the qubits in D’Wave’s QPUs. By qualitative and quantitative analysis, architects and engineers can create multiple optimal layouts which can be used to streamline the creation of digital doubles of the factory. The model created is flexible, as it takes no specialized data and just requires the rate of transportation of product between functional units in a pipeline, and the distance between each position with every other position. This makes this method be applicable to any sort of factory, as long as it follows pipeline based production style.