Anticipatory optimization for dynamic decision making relies on a number of different
scientific disciplines. On a general level, the foundations of the field may be
localized at the intersection of operations research, computer science and decision
theory. Closer inspection reveals the important role of branches such as simulation,
metaheuristics, Markov decision processes, dynamic programming and data mining.
Moreover, realization of an advanced approach to anticipatory optimization is hardly
possible without supporting database technology and without profound knowledge
about the problem domain under consideration.
However, bringing together all these elements is not an end in itself. Rather
the integral nature of anticipatory optimization reflects the complexity of dynamic
decision problems as they increasingly occur within various operational contexts.
Ultimately, the need for anticipatory optimization results from the fact that today
many companys’ operations are carried out under strongly dynamic and uncertain
conditions.
Starting from a discussion of origins and impacts of these conditions, the following
chapters elaborate on anticipatory optimization for dynamic decision making.
The presentation has three major facets. First of all, it focuses on consistent integration
of the methodological building blocks required for making an anticipatory
decision. Moreover, the presentation distinguishes between different degrees of anticipation
that may be realized by such a decision. Last but not least, an exemplary
class of dynamic decision problems from the field of vehicle routing is considered
for detailed investigation.
Going all the way from understanding the nature of dynamic decision making
to realization of anticipatory optimization at its best involves certain challenges.
Facing these challenges has been my everyday experience throughout the past years
of research. I am grateful for this experience as it comprises numerous fascinating
insights into the subject of dynamic decision making and beyond.