Information Technology Reference
In-Depth Information
Self-Organizing Logistics Process Control:
An Agent-Based Approach
Jan Ole Berndt
Center for Computing and Communication Technologies (TZI)
Universitat Bremen, Am Fallturm 1, 28359 Bremen, Germany
joberndt@tzi.de
Abstract. Logistics networks face the contradictory requirements of achieving
high operational effectiveness and efficiency while retaining the ability to adapt to
a changing environment. Changing customer demands and network participants
entering or leaving the system cause these dynamics and hamper the
collection of information which is necessary for efficient process control. Decen-
tralized approaches representing logistics entities by autonomous artificial agents
help coping with these challenges. Coordination of these agents is a fundamental
task which has to be addressed in order to enable successful logistics operations.
This paper presents a novel approach to self-organization for multiagent system
coordination. The approach avoids a priori assumptions regarding agent charac-
teristics by generating expectations solely based on observable behavior. It is for-
malized, implemented, and applied to a logistics network scenario. An empirical
evaluation shows its ability to approximate optimal supply network configura-
tions in logistics agent coordination.
1
Introduction
Logistics plays a major role in globalized economy. Industrial production and trade
require efficient and reliable supply networks. Growing interrelations between these
networks and the inherent dynamics of the logistics domain result in a high complex-
ity of global supply processes [9]. The application of conventional centralized planning
and control approaches to these processes suffers from that complexity. Therefore, de-
centralized methods become necessary which employ autonomous actors representing
logistics entities and objects [10].
From the artificial intelligence point of view, these autonomous entities can be repre-
sented by intelligent software agents to model logistics networks as multiagent systems
(MAS). These systems enable simulations, evaluations, and actual implementations of
new approaches in autonomous logistics [17].
In order to develop the aforementioned approaches, coordination and cooperation of
autonomous entities is a challenging task. In the logistics domain, coordination faces the
contradictory requirements of achieving high operational efficiency while retaining the
system's ability to adapt to a changing environment. On the one hand, supply networks
have to achieve high performance rates concerning asset utilization, cost reduction, and
customer satisfaction. On the other hand, they require flexible and robust structures in
order to react to unforeseen changes caused by the domain's inherent dynamics.
 
Search WWH ::




Custom Search