This chapter first shows how challenging conflict resolution is. The problem is combinatorial and cannot be solved with basic strategies. The chapter then describes different models that were imagined, namely centralized or distributed, and iterative or global. It focuses on resolution methods able to handle objective functions based on simulations. Metaheuristics are adapted to such functions because they do not require any properties compared to other methods. Next, the chapter details different approaches: a neural network (NN) trained by an evolutionary algorithms (EAs) for self‐separation, an EA and an ant colony optimization (ACO) algorithm for centralized global conflict resolution. Finally, it shows that one could define models that separate the problem definition from the resolution while still handling realistic uncertainties. This latter approach has the great advantage to allow comparisons between metaheuristics and exact optimization tools.
Conflict Detection and Resolution
2015-12-30
55 pages
Article/Chapter (Book)
Electronic Resource
English
How Important Is Conflict Detection to the Conflict Resolution Task?
Springer Verlag | 2016
|