Introduction to Jiuzhang Intelligent Logistics Products and Case Results
Keywords: smart logistics; supply chain optimization; route optimization; truck-cargo matching; independent research and development
Link on Zhihu (you can click the link directly here):
https://zhuanlan.zhihu.com/p/668143962?
Project goals
(1) Oriented to the integrated smart logistics optimization algorithm platform and ecosystem. An all-round intelligent algorithm customization system for prediction, planning, decision-making and data mining in transportation, warehousing, scheduling, allocation, network, after-sales, etc. in retail/express/delivery logistics and manufacturing supply chain management. It provides scientific support and technical guarantee to meet the needs of logistics business development and management decision-making under the new situation of integration/unmanned operation.
(2) Intelligent capacity/cargo matching platform. Compared with the existing vehicle-cargo matching platform, the support of intelligent optimization algorithms improves the efficiency of matching transportation demand and transportation resources, and reduces transportation costs and resource consumption.
core technology
(1) Independent core algorithm for large-scale, nonlinear, multi-objective, multi-level integer optimization and reinforcement learning. Compared with other logistics optimization solutions that mostly rely on foreign commercial operations optimization software (such as CPLEX, GUROBI, SCIP), for specific logistics optimization problems, our algorithm is more efficient (shown in Table 1) and is not affected by international trade disputes.
|
|
G
UROBI
|
Hybrid vehicle planning system |
|
Total mileage (km) |
1
08328
|
2
1093(
↓80.5%
)
|
|
Total number of vehicles (vehicles) |
366
|
212（
↓4
0.8%
）
|
|
Calculation time (hours) |
5
|
0
.75(
↓8
5%
)
|
Table 1: Comparison of solution efficiency between simulated data and standard integer programming problems
(2) Logistics intelligent optimization algorithm customization platform. Through meta-learning and super-heuristic technology, based on constraints, goals and big data, intelligent algorithms that can solve logistics decision-making optimization problems and can be self-learning and self-evolving are automatically generated for specific enterprises and specific businesses.
Typical cases
Algorithm development of an automotive parts factory scheduling optimization system
(1) Project Introduction: Entrusted by an automobile group company, through an industry-university-research cooperation project (see Table 2), deep learning and heuristic optimization algorithms were developed for the logistics scenarios of parts entering the factory and vehicle distribution.
(2) Project results. The new system developed based on the subject algorithm has been deployed in actual business:
Algorithm library that fully supports transportation path optimization, loading optimization and queuing optimization.
Support large-scale (100,000 orders or 10 million kilometers per day) logistics scheduling needs in real time.
Theoretical calculations can reduce costs by about 10% (annual logistics costs exceed 10 billion yuan).
|
|
Existing system |
new system |
|
Total number of journeys (km) |
34712
|
32605(
↓6%
)
|
|
Total number of vehicles |
113
|
100
(
↓11.9%
)
|
|
Average scheduling time |
42 person hours |
15 people minutes (
↓99.4%
)
|
Table 2: Efficiency comparison
some cases
|
Typical customer
|
scene
|
Plan
|
value
|
|
An automobile company |
Parts entry and exit plan 500 supplier sites, 1800 retail stores, 7 warehouses, 17 professional centers |
Path planning for multiple vehicles, multiple goods, and alternate loading and unloading Loading optimization, queuing optimization |
Daily average
100,000 orders travel
10 million kilometers annual savings
6-1 billion |
|
A transportation company |
convoy
56 vehicles, transportation positioning and planning and dispatching |
Urban transportation and distribution Loading optimization |
Save planning time
80%
|
|
An autonomous vehicle company |
Unmanned vehicle park dispatch management Ordering and delivery services |
Intelligent scheduling, route optimization, and delivery app |
Unmanned vehicles successfully landed in the park |
|
A logistics park |
Intermodal transport by trucks and trains |
Multimodal transportation intelligent dispatch planning
|
Successfully launched and supported every day
30+ truck dispatching and train schedules |
|
An aviation logistics company |
Realize door-to-door logistics integrated transformation business 860+ routes |
Road network topology optimization Door-to-port, port-to-port, and port-to-door scheduling optimization |
Successfully launched, the card transfer business has completed the intelligent booking business, and the operation is being optimized. |
|
A logistics information technology company |
combine
TMS system information for path optimization and scheduling optimization |
Algorithm as a service, algorithm module embedding
In TMS system |
Help enterprises improve their intelligence level and market competitiveness |
|
A food company |
3D boxing, every day
50,000 cartons, each box to contain different orders |
Quickly find the most suitable cartons |
Saves more than labor on average
30% carton material |
|
A manufacturing company |
60 warehouses and manufacturing centers realize global optimization of the transportation of raw materials and finished products |
Warehouse network optimization, transportation scheduling optimization |
implementation cost
5-10% reduction
|
|
A new energy vehicle operating company |
300 new energy vehicles, charging combined with freight dispatching plan |
Transportation scheduling plan combined with charging scheduling |
save
5-12% cost Improve the utilization rate of charging piles Reduce the purchase of new energy vehicles |
Table 3: Some cases
Demo and introduction materials:
(1) Introduction to the Demo website:
Links:
https://pan.baidu.com/s/1nC9anK8w0gEWpMhMi_rOPA?pwd=2023
(2) Introduction to a certain automobile logistics company:
Links:
https://pan.baidu.com/s/1cd4cKtWiMvBp3LQ8BWdgmQ?pwd=2023
(3) Demo website:
http://47.103.101.234:8000/
Username/password: demo08/%xaauq
(4) Intelligent decision-making documents:
https://mp.weixin.qq.com/mp/homepage?__biz=MzA3MzM3NDIyNw==&hid=1&sn=65b32a9b567e4217a964881425aa6635&scene=1&devicetype=Windows+10+x64&version=6309071d&lang=zh_CN&nettype=WIFI&ascene=0&session_us=gh_690650ba0de6&uin=&key
Machine-translated from the Chinese original 久章智能产品与案例介绍 (published 2023-11-22). WeChat original ↗