European Journal of Scientific Research - EuroJournals

765 Rabindra Ku. Jena, Musbah M. Aqel, Gopal K. Sharma and Prabhat K. Mahanti
continues with solutions of the second non-dominated front, followed by the third non-dominated front
and so on. Since the overall population size of Rt is 2N, not all fronts may be accommodated in N slots
available in the new population. All fronts which could not be accommodated are simply deleted.
When the last allowed front is being considered, there may exist more solutions in the last front than
the remaining slots in the new population. This scenario is illustrated in Figure 4. Instead of arbitrarily
discarding some members of the last front, it would be wise to use a niching strategy to choose the
members of the last front, which reside in the least crowded region in that front. But we will not go into
too much detail on that. For more details refer to (Deb, 2002).
6. Problem Formulation
6.1. Basic Idea
Like other algorithms in area of design automation, the algorithm of NoC communication architecture
is a hard problem. Our attempt is to develop an algorithm that can give near optimal solution within
reasonable time. Genetic algorithms have shown the potential to achieve the dual goal quite well (Jena
et al, 2006; Lahiri et al, 2000; Srinivasan and Chatha, 2005).
As shown in Figure 5 and discussed in section 1, the problem is solved in two phases. The first
phase (P-I) is basically a task assignment problem (TA-GA). The input to the problem is a TG. We
assume that all the edge delays are a constant and equal to Average Edge Delay (AED) (Banerjee et al,
2004).The output of the first phase is a Core Communication Graph (CCG).
The main task of the second phase (P-II) is communication synthesis which basically deals with
mapping of cores to the tiles of the NoC backbone using multi objective genetic algorithm. The next
section discusses each phases in detail.
Figure 5: An overall design flow.
TA-GA
CTA-GA
CTS-GA
NoC Architecture Description
Task Graph
Task Assignment
Evaluation
Core Communication Graph
Core Tile Mapping
Evaluation
Core Tile Switch Mapping
Evaluation
Optimize NoC
6.1.1. Task Assignment Problem (TA-GA)
Given a task graph TG with all edge delay are constant and equal to average edge delay and IPs with
specifications matrix containing cost and computational energy. The main objectives of this phase are
to assign the tasks from the task graph to the available IPs in order to:
(i) minimize the computational energy by reducing the power consumption.
P-I
P-II