NEW TECHNOLOGIES IN CONTAINER TERMINAL AUTOMATION AND LOGISTICS
The development of increasingly sophisticated algorithms and logic, among which those referred to AI, makes it possible to delegate to machines decision-making tasks hitherto left to people.
Among the countless possibilities given today to further digitize all operational phases of container terminal management thanks to the advance of new technologies, we limit ourselves to pointing out those due to new algorithms applied to operations and closely related to logistics aspects. So, container handling and positioning in all its complexity and with respect to time, cost and resource parameters that determine its effectiveness and operational efficiency. It is of recent evidence how the development of increasingly sophisticated algorithms and logic, among which those referred to as Artificial Intelligence (AI) are in vogue, makes it possible to delegate to machines decision making tasks hitherto left to people. It is therefore on this issue that the following considerations are developed. To provide some clarity, let us try to summarize the main possible cases of new technologies that can be seen applied in the area aforementioned to achieve clear improvements.
POSITIONING AND MOVEMENT USING DETERMINISTIC LOGIC
Principles: It is logically possible to determine the position of a means of movement, according to rules that seek to minimize its time and cost of use and maximize its yields. Reasoning of this kind is accomplished by identifying deterministic formulas and algorithms; they are essentially the domain of operations research, which has also found wide application in container logistics and marine terminals. Clearly this science, while solidly established, is subject to continuous improvement by providing new tools, new equations and new reasoning. The general algorithm is based on solving a number of equations with an equal number of unknowns that cannot be fixed to a certain value by the operator. In essence, it is the area of classical optimization.
Areas of application: In general, the application of these techniques is suitable for relatively simple cases where the unknowns are not many and so are the applicable rules, which, moreover, are definable in a certain way. The majority of TOS on the market today make use of this technology to support the decisions of the operator who fixes positions, orders movements, and establishes equipment and people operating on the movements. To exemplify to the fullest extent, let us think of the movement and loading and unloading of a container whose location is determined on the ship, and whose position is to be determined on the yard (in comparison with all other objects in the area of interest). In this case, the choice of operating means and the location of the container are done by minimizing time and cost and are thus related to the ship's positions on the yard and the available space and means. Evidently this is usually a deterministic problem and fixed by very precise equations. Reality however, is far more complex and different, and deterministic laws can only approximate a result – and in any case become very complicated by multiplying equations and variables.
ANALYSIS OF LARGE MASSES OF DATA (AND CASES) AND LEARNING FROM EXPERIENCE
Principles: A step further toward more refined information management is achieved by evaluating experience originated from facts. Facts that can be inferred from movements and operations at the dock by ship and on containers. It is a treatment of large masses of data (big data) that can be processed statistically, from which stable behaviors can be inferred over time through rules that are not obvious and not mathematically translatable (machine learning), according to the operations research in the previous paragraph. This is the field of predictive analysis.
Areas of application: The application areas of this technology are oriented in the field of decision-support tools, proposing movement and workload optimization solutions for complex management areas such as positioning over an entire yard of containers or on a quay regarding ship approaches
INTELLIGENT AGENTS CAPABLE OF LEARNING
Principles: A more elaborate way still to indicate the optimal behavior of sets of objects operating independently but related with defined goals and an environment that is also definable (at least in its behavior), is to assign and design intelligent agents for each object in question. That is, a system of elements that perceives its environment through sensors and acts through actuators according to logic fixed by its algorithms that establish perception, autonomy, action and learning. With differences due to being model based agents, goals or utilities. This is the domain of intelligent operations.
Areas of application: Such systems, in a very complex area such as the terminal, are preconceived but still difficult to implement. They may find application in the construction of a system that acts completely autonomously by regulating itself and then proposing a fully automated operation. They could represent the way to automate in the future a complex area such as the whole terminal (definable neither as a simulation nor as an emulation, but as a real replacement of the operation with a fully automatic system). A classic application of AI for a newly developed TOS. More narrowly, intelligent agents could be applied to circumscribed areas such as a series of vehicles thus endowed with intelligence and consequent autonomy.
ABOUT THE COMPANY: Data and System Planning (DSP) is based in Switzerland, founded in 1986. DSP provides worldwide IT solutions and business operation consultancy taking advantage of its specialized and proven know how in informatics technologies applied to the shipping industry, port and terminal management and intermodal transportation. DSP offers a large portfolio of professional services and products to support terminal operations processes and systems. Its expertise covers planning and control optimization, training, testing, process automation design and integration, reporting and business intelligence, billing and tariff management systems. DSP is a NAVIS partner since 2007 covering all implementation, maintenance and personalization services.
ABOUT THE AUTHOR: Giambattista Ravano obtained his degree in Physics from the Federal Institute of Technology Zurich (ETHZ) where he then worked as a teacher and research assistant. After a professional specialization in Information Technology and a long and familiar experience in the maritime industry, he founded DSP. Now Giambattista is leading the company as President, with special assignment on software engineering, security, innovation and sustainability issues. He has long been Professor of Informatics in Switzerland.
“A STEP FURTHER TOWARD MORE REFINED INFORMATION MANAGEMENT IS ACHIEVED BY EVALUATING EXPERIENCE ORIGINATED FROM FACTS. [...] THIS IS THE FIELD OF PREDICTIVE ANALYSIS.”