Title :
A neural net based self organising scheduling algorithm
Author :
Hemani, Ahmed ; Postula, Adam
Author_Institution :
Swedish Inst. of Microelectron., Kista, Sweden
Abstract :
Scheduling is a crucial task in behavioural synthesis and a NP-hard optimisation problem. Neural net computation paradigms bring potential for efficient solutions to such problems. This paper presents a new scheduling algorithm based on Kohonen´s rule for self organisation. The algorithm has an inherent hill climbing mechanism, copes with a comprehensive set of constraints and can be implemented on massively parallel structures. Its performance on well known benchmark examples, presented in the paper, is on par with the best reported
Keywords :
circuit layout CAD; neural nets; scheduling; self-organising storage; Kohonen´s rule; NP-hard optimisation; behavioural synthesis; benchmark; neural net based self organising scheduling algorithm; Costs; Hardware; High level synthesis; Microelectronics; Neural networks; Pattern recognition; Processor scheduling; Scheduling algorithm; Surges; Traveling salesman problems;
Conference_Titel :
Design Automation Conference, 1990., EDAC. Proceedings of the European
Conference_Location :
Glasgow
Print_ISBN :
0-8186-2024-2
DOI :
10.1109/EDAC.1990.136634
