Covering Disruptive Technology Powering Business in The Digital Age

Fujitsu Limited and Toyota Systems Corporation successfully demonstrated optimisation of supply chain and logistics network operations essential to support automobile production in a joint trial leveraging Fujitsu’s Quantum-Inspired Digital Annealer computing solution.

The Digital Annealer solves complex combinatorial optimisation problems at speeds not possible with current conventional computing technologies, and was used to rapidly calculate variables including the number of transport trucks, total mileage, and package sorting tasks, determining the most cost-effective approach for an automotive parts supply chain optimisation problem with more than 3 million possible delivery routes to dozens of factories.

Ultimately, within 30 minutes, the trial revealed it was possible that an optimal route was calculated that can potentially reduce logistics costs by approximately 2 to 5% by discovering effective, previously unidentified distribution routes, thereby improving loading efficiency and streamlining transportation-related expenses.

Fujitsu and Toyota Systems will further verify and refine the solution in the future, aiming to commercialise it for real-world supply chain and logistics operations.

Outline of the Field Trial

Fujitsu and Toyota Systems endeavoured to find a solution through joint trials and the development of innovative technologies with these challenges in mind.

Toyota Systems was first established in January 2019 as an IT solutions company to support Toyota Motor Corporation and its group companies in the development of technology offerings. Toyota Systems has been conducting research with Fujitsu on the use of quantum computing since prior to the establishment of the company.

Toyota Systems and Fujitsu applied Fujitsu’s Digital Annealer quantum-inspired technology during the trial to formulate and solve an optimisation problem involving a large-scale logistics network with more than 3 million route candidates. The objective of the trial was to quickly determine the route with the lowest distribution cost from millions of potential candidates for procuring parts from hundreds of suppliers and delivering them through several transit warehouses to dozens of factories. Distribution costs were calculated based on variables including the number of trucks, total travel distance, and the amount of work done in sorting packaged parts.

In order to apply this technology to real-world operations, a new system was developed that incorporates a large-scale problem-solving technology previously developed by Fujitsu Limited and Fujitsu Laboratories, Ltd., and the following two technologies:
– Global Search Technology: searches a wide range of possible solutions while efficiently escaping from a local solution groups
– Dynamic Multipoint Search Technology: dynamically determines search starting points based on the current search status and interim results

(0)(0)