A.M. Douma

Design and evaluation of a simulation game to introduce a Multi-Agent system for barge handling in a seaport

Multi-Agent systems have been studied extensively, but only a few of these systems are deployed in practice. Essential to get a system implemented is acceptance. In a distributed setting this is challenging, especially when one deals with multiple independent and competing companies. We share our experiences with the use of a real-time multi-player simulation game that we developed to illustrate a Multi-Agent system for the barge handling problem in the Port of Rotterdam. We experienced that the game has many advantages over a more passive approach, such as vocal presentations. We conclude that the game has considerably contributed to the acceptance of the illustrated Multi-Agent system.

Aligning the operations of barges and terminals through distributed planning

In this thesis we consider the barge handling problem, which is about the alignment of barge and terminal operations in a port. Complicating factor in tackling this problem is that centralized planning is not acceptable for the parties concerned. In our research we explore an alternative: distributed planning. We develop a Multi-Agent system (MAS) consisting of barge operator agents and terminal operator agents. We propose a specific interaction protocol based on service-time profiles through which barge and terminal operators are able to improve their planning. To evaluate the performance of our MAS we conduct simulation studies. In this way we get insight in the performance of the system as a whole as the result of the interactions of barge and terminal operators. Additionally, we compare the performance with an off-line benchmark, resembling central coordination. Our findings are promising. Our interaction protocol supports an efficient negotiation between barge and terminal operators. Our MAS allows for real-time alignment of barge and terminal operations such that the overall system performance is well within the range of central planning performance. Moreover, our MAS suppresses the propagation of disruptions, making the operations of barges and terminals more reliable. Experiences with the management game that we designed to explain our solution to practitioners, are encouraging and indicate that our MAS may be acceptable for barge and terminal operators and implementable in practice. Throughout the thesis we took the Port of Rotterdam as our source of inspiration, although our model is applicable to general multi-terminal, multi-barge settings. Generally we conclude that our Multi-Agent system is a promising solution for the barge handling problem and can result in a significant improvement in practice.

Degrees of terminal cooperativeness and the efficiency of the barge handling process

We study the effect of different degrees of cooperativeness on the efficiency of the barge handling process. The barge handling process is about the alignment of barge and terminal operations in a port. The efficiency of this alignment process depends on the cooperation of different players and especially terminal operators. In an earlier study we developed a Multi-Agent system assuming that terminals are fully cooperative, i.e., they provide insight in their occupation and make reliable appointments. In practice, terminals might decide to be less cooperative. We consider three degrees of cooperativeness: low, partly, and full cooperativeness. Experimental results indicate that there are two reasonable alternatives: full and low cooperativeness. In the lowly cooperative case, the lack of cooperativeness of terminals is compensated by cooperation among barge operators. We provide an extensive discussion on both alternatives. Our results provide useful insights for barge and terminal operators in the options they have to improve the barge handling process.

Agent-Based Support for Container Terminals to Make Appointments with Barges

We consider a container terminal that has to make appointments with barges dynamically with only limited knowledge about future arriving barges, and in the view of uncertainty and disturbances. We study this problem using a case study at the Port of Rotterdam, considering a proposed multi-agent system for aligning barge rotations and terminal quay schedules. We take the perspective of a single terminal participating in this system and focus on the decision making capabilities of its intelligent agent. Using simulation, with input settings based on characteristics of the larger terminals within the Port of Rotterdam, we analyze the benefits of our approach. We conclude that a terminal can increase its utilization significantly by using various sources of flexibility in the operational planning.