Independent research institute for optimization AI BSAI announced that the AGV group control algorithm “PYUTHIA” can be applied to an automobile manufacturer’s distribution warehouse, resulting in a maximum efficiency improvement of 1.5 times.

Smith & Motors Co., Ltd.
[Independent research institute for optimization AI] BSAI announced that the AGV group control algorithm “PYUTHIA” can be applied to an automobile manufacturer’s distribution warehouse, resulting in a maximum efficiency improvement of 1.5 times.

Black Stone Algorithm Institute (Hongo, Tokyo, hereafter: BSAI) has installed “PYUTHIA”, which optimally controls several to 1,000 AGVs in an environment where multiple AGVs work together, in an automobile manufacturer’s distribution warehouse. As a result of applying this technology to a vehicle, the company announced that it is possible to improve efficiency by 1.5 times.
A video that specifically compares the case of optimization using PYUTHIA and the case of not, visually showing the degree of efficiency when actually introducing PYUTHIA.
[Image 1

Issues of AGV group control in the manufacturing and logistics industries In the manufacturing and logistics industries, labor shortages due to population decline are a serious issue, and automation and efficiency improvements using robots and AI are required. Among them, AGV (* 1) is the key to automating the logistics in the hall.
Logistics automation in large-scale factories requires coordinated operation of multiple AGVs, which requires advanced group control algorithms (*2).
However, in the conventional AGV group control
1. The routes of AGVs collide due to the lack of proper coordination commands. 2. There are restrictions on the number of units that can be controlled 3. Failure to minimize path length, causing AGV to take detours 4. A plan that matches the actual speed and acceleration of the AGV cannot be created, and an unrealizable plan is output.
There was a problem.
*1 Abbreviation for Automated Guided Vehicle, meaning “automated guided vehicle (or unmanned guided vehicle)”.
(Representative example: AGV that transports each shelf installed in Amazon’s distribution warehouse)
*2 Technology to operate multiple robots simultaneously in a factory/warehouse An example of 3. is when AGVs are approaching from both directions into a passage where only one AGV can pass. In the conventional method, the AGVs approach each other until just before the collision, and to avoid the collision, it is necessary to make an emergency stop and re-plan the route.
[Image 2

Delay caused by conventional group control algorithm (right in image) and control of PYUTHIA (left in image)
The disadvantages of these challenges are
・Increase in cost due to increase in the number of AGVs introduced ・Takt time delays due to unexpected detours at the time of route planning ・Since it is not possible to control more than 100 AGVs at the same time, it is impossible to introduce it on a very large scale. and so on.
In particular, the second point is a fatal disadvantage linked to the observance of the delivery date. If the algorithm cannot solve the problem, the on-site operations must flexibly respond, which places a heavy burden on the on-site operations.
“PYUTHIA” solves these problems with cutting-edge academia technology. Comparison of conventional system and PYUTHIA
PYUTHIA achieves up to 1.5 times efficiency compared to conventional systems. The video below will explain in detail.
Movie 1: Group control by conventional system
[Video 3: https://prtimes.jp/api/movieim.php?url=www.youtube.com/watch?v=lfsg81EsZNc] Group control by conventional systems
1. Each time a collision is about to occur, the route is re-planned, which causes a detour and a loss of time.
2. Since the actual time required for each task such as “lifting, lowering, transporting” when AGV transports is not considered, it is necessary to estimate the buffer for the pre- and post-process, resulting in time loss.
There are wastes such as this, and therefore we have not been able to pursue efficiency with AGV.
Video 2: Group control using PYUTHIA
[Video 4: https://prtimes.jp/api/movieim.php?url=www.youtube.com/watch?v=yJH06aPekXQ] Compared to Movie 1, group control using PYUTHIA
1. Since the route is planned considering how other AGVs will move in the future, collisions are almost non-existent.
2. “Speed ​​when AGV is transporting/not transporting things” “How long it takes to pick up/drop an item”
“Time it takes for an AGV to turn 90 degrees/180 degrees”
Consider the time it takes to actually perform the task. This makes it possible to accurately estimate the transportation time, so there is no need to estimate a large buffer for the processes before and after transportation by AGV.
Thoroughly eliminate waste such as AGV and make ultimate efficiency improvement possible.
In fact, it has become clear through verification that a maximum efficiency improvement of 1.5 times can be expected in a simulation that is introduced at an automobile manufacturer’s distribution warehouse.
Assumed introduction case of “PYUTHIA”
1. Factories and warehouses that have already introduced more than one AGV 2. AGV manufacturers
At sites where multiple AGVs have already been introduced, even though they are already benefiting from the automation of transportation, further efficiency improvement is an issue. Optimal control of AGV by PYUTHIA makes it possible to realize more efficient factories and warehouses.
In addition, we are considering cooperation with manufacturers that introduce AGVs to factories and warehouses such as those mentioned above, in the form of installing PYUTHIA in the AGVs they sell and research and development for further efficiency.
Future prospects
In the future, we will promote further research and development of our unique AGV group control algorithm “PYUTHIA” and provide more efficient systems to the manufacturing industry, logistics industry, and AGV manufacturers.
In addition, BSAI is researching and developing state-of-the-art algorithms such as the scheduler “APOLON” that realizes optimal operation plans for machines and equipment at manufacturing and logistics sites. Through this, we aim to “realize an optimal world with state-of-the-art algorithms”.
About Black Stone Algorithm Institute
[Image 3

The Black Stone Algorithm Institute is an independent research and development group whose mission is to achieve “realization of an optimal world with state-of-the-art algorithms”.
-Press release from Black Stone Algorithm Institute-
・[AI venture from the University of Tokyo] Established BSAI, an independent research institute, and started providing “PYUTHIA”, a unique technology that realizes AGV group control of several to 1,000 units.
https://prtimes.jp/main/html/rd/p/000000023.000066955.html
■ Inquiries regarding this press release
Name: Takuma Domoto
Email address: t-domoto@trustsmith.net
Phone number: 03-6822-3276
Company HP: https://www.trustsmith.net/
Twitter: https://twitter.com/TS_Domoto_42
Facebook: https://www.facebook.com/takumadomotoku

Details about this release:
https://prtimes.jp/main/html/rd/p/000000008.000071107.html