A proof-of-concept – PoC test of a proprietary Yokogawa AI solution for automating outbound shipment loading planning for Hokuetsu Logistics Corporation, a subsidiary of Hokuetsu Corporation, was successfully completed by Yokogawa Electric Corporation’s subsidiary, Yokogawa Digital Corporation.
This Yokogawa AI solution can automate a loading planning process that was previously completed by multiple experts, replicating expert-level decision making processes. This significantly reduces the time needed for planning and optimizes loading efficiency based on shipping regions and paper product shapes.
In July 2025, Hokuetsu Corporation and Hokuetsu Logistics Corporation made the decision to formally accept and start utilizing this AI technology based on the outcomes of this proof of concept test.
Background Information
The primary operations of Hokuetsu Corporation are the production and distribution of pulp and paper products. The shape of the products, the vehicles used for transportation, and particular criteria at the export destination are some of the limitations that apply to the outbound shipment of paper and pulp products.
It’s also critical to lessen the workload for truck drivers by confining shipments to single destinations whenever feasible and, in cases when several destinations are required, scheduling them to be as close to one another as feasible.
Developing loading plans becomes more challenging with traditional combinatorial optimization techniques due to the complexity of the objectives and constraints, and the required computations take longer to finish.
Both the training of loading planning experts and attempts to automate the loading planning process have been challenging.
Development of AI Solution and Implementation of PoC Test
With the aim of developing an AI solution that could automate this loading planning process, Yokogawa’s AI consultants interviewed Hokuetsu Corporation’s highly skilled planning specialists to understand their thinking processes and gain a clear understanding of site operations.
Based on these insights, Yokogawa developed a proprietary AI-driven loading planning solution that replicated the decision-making expertise of specialists. A PoC test was then implemented that confirmed the following outcomes.
- Quick completion of loading planning
It was confirmed that the AI loading planning tool is capable of generating accurate loading plans, in a significantly reduced timeframe of less than 10 seconds for individual planning, even while taking into account complex conditions such as shipped items, vehicles, and specific requirements at delivery destinations.
- Successful incorporation of expert-level capability to reduce burden on delivery personnel
It was also confirmed that the planning tool mimics the capabilities of specialists to decrease the burden on personnel by consolidating the number of delivery destinations to one or two. When multiple arrival points are required, the system prioritizes the nearest destinations and generates plans that also take load capacity into account.
Future Plans
Yokogawa will contribute to the improvement of operational efficiency and the construction of a sustainable operation system at the Group’s logistics sites by providing an AI loading planning solution that factors in individual circumstances at each company and mimics the reasoning of expert personnel.
