| Full text | |
| Author(s): Show less - |
Silva, Cristiano M.
;
Sumika, Fernanda
;
Guidoni, Daniel Ludovico
;
Lopes Santos, Leonardo Alvarenga
;
Machry Sarubbi, Joao Fernando
;
Mokhtari, Somayeh
;
Rodrigues Santos, Matheus F.
;
Silva Nogueira, Jose Marcos
;
Varga, P
;
Granville, LZ
;
Galis, A
;
Godor, I
;
Limam, N
;
Chemouil, P
;
Francois, J
;
Pahl, M
Total Authors: 16
|
| Document type: | Journal article |
| Source: | PROCEEDINGS OF THE IEEE/IFIP NETWORK OPERATIONS AND MANAGEMENT SYMPOSIUM 2022; v. N/A, p. 4-pg., 2022-01-01. |
| Abstract | |
In this work, we overview a novel strategy that can support the decision making of systems dealing with combinatorial problems. We are particularly interested in supporting decision making in nearly real time systems. To target this goal, the proposed strategy must have a configurable lime threshold to output the solution. Obviously, the solution presented will not be the optimal one because we are dealing with NP-complete problems. However, the evolution of this research may lead to an interesting class of solutions specifically designed for the real time scenario, where machines have to learn fast how to solve a problem, and lake the appropriate measures to avoid a catastrophic situation (such as car crashes, course corrections, and general automation systems). The initial experiments show interesting results, with high gains when compared to similar approaches in the literature. (AU) | |
| FAPESP's process: | 20/05126-6 - Algorithms and protocols for intelligent transportation systems |
| Grantee: | Daniel Ludovico Guidoni |
| Support Opportunities: | Regular Research Grants |