Advancements in Dynamic Control of Temporal Constraint Networks

Seminar for the Admission to the Final Exam

October 5, 2018
Versione stampabile

Luogo: PovoZero, via Sommarive, 14 - Povo - Sala Seminari "-1"
Ore: 12:00


  • Cairo Massimo (PhD Student in Mathematics)


Temporal Constraint Networks are models used to represent knowledge about the time when some events occur, and to perform automatic reasoning about these occurrence times (Automated Temporal Reasoning).
These models can be also used to represent restrictions on the execution time of some actions, and to automate the scheduling of these actions, in order to satisfy all the temporal restrictions (Automated Scheduling).
When the models include uncertainty, for example when the duration of some actions is not under the control of the agent, then the problem of Automated Scheduling becomes a problem of Dynamic Control, meaning that the executing agent needs a strategy to execute the actions, and the scheduling can change dynamically according to the observed behavior of the environment (such as the actual duration of the uncontrolled actions).
We present some recent results regarding Dynamic Control for two of the main Temporal Constraint Network models which include uncertainty, namely Simple Temporal Networks with Uncertainty (STNUs) and Conditional Simple Temporal Networks (CSTNs). These results include: algorithms and complexity hardness results for dynamic controllability checking, systems of constraint propagation rules, theorems of equivalence between variants of models, algorithms for real-time execution.

Referente: Rizzi Romeo