Skip to main content



  1. Home >
  2. Feasibility Study – Utilization of Artificial Intelligence and Big Data Analysis to Alleviate Port Congestion

Feasibility Study – Utilization of Artificial Intelligence and Big Data Analysis to Alleviate Port Congestion

Fumitaka SATO Senior Consultant
Atsushi IKEGAMI Consultant

Thursday, March 28 , 2019

Feasibility Study through the cutting edge artificial technology and big data analysis

Fujitsu Limited and Fujitsu Research Institute, partnered with the Agency for Science, Technology and Research (A*STAR) and the Singapore Management University (SMU), as the Center of Excellence (CoE) based on a Master Research Collaboration Agreement (MRCA) signed in 2014, completed the feasibility study program in collaboration with an international port in Singapore in order to contrive solutions, by utilizing artificial intelligence and big data analysis, for ‘Berth(*1) Congestion’ and ‘Gate Congestion’ problems which Singapore has been facing for long as a country with the high volume of maritime trade and traffic.

Problem Descriptions
Berth Congestion - Background and Problems

Our partner for the feasibility study – the port in Singapore which mainly handles bulk cargos(*2) – once finalized berth plans(*3) only a day before the arrival of vessels due to the ‘first come and first served (FCFS) ’ practice. This FCFS practice caused vessels approaching the port to rush and wait usually for several days off the coast before they berth.

After the arrival of vessels, they stayed on average for four (4) days to unload cargos. However, since bulk cargos are more likely to be affected in its nature by weather, once vessels encountered negative weather conditions, it forced them to change their original plans and stay longer for unloading, having caused the port to readjust crane operations, human resource allocations, and berth plans for vessels following one after another. In addition, the port heavily relied on skilled employees to handle this situation which implies their performance differs from time to time and hence it also became a major obstacle to improve the efficiency of the management of their operational environment.

Figure 1: Berth Congestion


Gate Congestion – Background and Problems

With further economic growth in Singapore projected, it is expected that the volume of cargos would constantly increase. As a result, it could cause a heavy traffic congestion during a peak time in specific areas where port gates are located, since the number of trucks with cargos, which come and go based on instructions from consignors, would increase regardless of schedules of vessel arrival.

In addition, since the port relied on skilled employees to adjust and manage opening and closing hours for gates/lanes, it became a major obstacle to optimize the allocation of employees in charge of gate operations.

Figure 2: Gate Congestion


Our Solutions
Our Approach to Problem Solving

The CoE team (IHPC, SMU, Fujitsu, and Fujitsu Research Institute), which has strong advantages in the field of maritime supply chains, mathematical optimization, and future trend forecasting, received mass data related to berth planning and gate management operations with the support of the port, having promoted the improvement of congestion at the port through statistical analyses of a variety of trends and field demonstrations.

Optimization of Berth Booking

Based on the results of our statistical analyses, the CoE team contrived the advanced booking system with optimization functions: the system receives request for berthing from 31 to 60 days before their arrival and thus can more easily take control of the number of vessels coming. In addition, bookings made through the system are optimized mathematically, maximizing the berth utilization rate (abbreviated as BOA rate) without changing berthing requests from vessels as much as possible. This enables the port to improve the berth utilization as well as the quality of service for vessels.

Figure 3: Optimization of Berth Plans


Optimization of Gate Operations

We built a traffic volume / congestion model from past traffic volume performance data and forecasted traffic trends. Based on this model which involved discrete event simulation, we formulated more efficient operational plans by predicting the change in traffic volume in response to the change in cargo volume.

Figure 4: Simulation of Traffic Congestion Prediction


For the question ’How can we solve traffic congestion?’ which is commonly asked however often leads us to a labyrinth of logics, we found the optimal answer utilizing mathematical optimization technology. If we only focus on solving traffic congestion, it could be done by opening gates all the time, or physically increasing the number of gates. What is unique in our approach is that the optimization engine proposes operational plans which alleviate congestion at gates as well as minimize the cost of operations.

Figure 5: Optimization of Gate Operations



The following outcomes were gained as a result of our feasibility study program.

  • Berth Congestion
    With our method and optimization technology, optimized berth plans, which cater to berthing requests from vessels to the maximum extent, were created, resulting in improving the BoA rate, decreasing waiting times and leveling the berth utilization rate.
  • Gate Congestion
    With our method and optimization technology, optimized gate plans alleviated congestion and saved the operational time by 17% (on average 9%).


(*1) ‘Berth’: A designated location in a port or harbor used for mooring vessels when they are not at sea. Berths provide a vertical front which allows safe and secure mooring that can then facilitate the unloading or loading of cargo or people from vessels
(*2) ’Bulk Cargo’: Goods such as coal, grains, oil, or chemicals that are not packaged in any type of container, and are transported in large quantities
(*3) ’Berth Plan’:Plans of loading, unloading and other related operations with time slots for each vessel.
(*4) ’Maritime Supply Chain’:Chain of overall value from suppliers to the end user in maritime logistics
(*5) ’BoA rate’: An abbreviation of ‘Berth on Arrival rate’ which is the ratio of requested dates for berthing to actual berthing dates.

*Note: The figures, company names and proper nouns mentioned in this example are as of the date of publication, and may change over time.

Related Trends

Consulting Service with AI technologies(Japanese)

Related Service

Optimization for Berth booking operations
Optimization for Gate Operations

佐藤 文孝(さとう ふみたか)

Fumitaka SATO (佐藤 文孝)/ Senior Consultant /Fujitsu Research Institute
Leads to solve various problems by data analysis from the viewpoint of data scientist.

池上 敦士

Atsushi IKEGAMI (池上 敦士)/ Consultant /Fujitsu Research Institute
Supports clients both in the public and private sector on policy planning and execution, particularly specialized in the fields of social infrastructure, international trade and national security. Atsushi also serves as a visiting fellow at Defense Technology Foundation and contributes to Defense Technology Journal.

list of 2019