Abstract

AbstractOptimized task scheduling is one of the most important challenges in parallel and distributed systems. In such architectures during compile step, each program is decomposed into the smaller segments so-called tasks. Tasks of a program may be dependent; some tasks may need data generated by the others to start. To formulate the problem, precedence constraints, required execution times of tasks, and communication costs among them are modeled using a directed acyclic graph (DAG) named task-graph. The tasks must be assigned to a predefined number of processors in such a way that the program completion time is minimized, and the precedence constraints are preserved. It is well known to be NP-hard in general form and most restricted cases; therefore, a number of heuristic and meta-heuristic approaches have so far been proposed in the literature to find near-optimum solutions for this problem. We believe that ant colony optimization (ACO) is one of the best methods to cope with such kind of problems pres...

Highlights

  • Today, the utilization of multiprocessor systems has been increased due to increase in time complexity of the programs and decrease in hardware costs

  • Some tasks need the data generated by the other tasks; the problem can be modeled using a directed acyclic graph (DAG) so-called task graph

  • Since the_total_number_of_the_ants is a constant initiated to 2500, for the big-enough numbers of n, we can assume that overall time-complexity of the proposed approach in sequencing subproblem belongs to the O(n2), which is equal or better than the traditional preintroduced heuristic methods

Read more

Summary

Introduction

The utilization of multiprocessor systems has been increased due to increase in time complexity of the programs and decrease in hardware costs. The BNP (Bounded Number of Processors) methods such as HLFET14 [17], ISH15 [5], CLANS16 [18], LAST17 [19], EFT18 [20], DLS19 [21], and MCP20 [14], in which the number of processors are restricted using list-scheduling technique (the proposed approach is in this group). They make a list of ready-tasks at each stage, and assign them some priorities.

Multiprocessor Task Scheduling
The HLFET Algorithm
The MCP Algorithm
The DLS Algorithm
The ETF Algorithm
Ant Colony Optimization
The Proposed Approach
Implementation Details and Experimental Results
The Utilized Dataset
The Experiments and Results
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call