A Hybrid Chemical Reaction Optimization Scheme for Task Scheduling on Heterogeneous Computing Systems
read more
Citations
Ensemble particle swarm optimizer
Bi-objective workflow scheduling of the energy consumption and reliability in heterogeneous computing systems
Use of Organic and Copper-Based Nanoparticles on the Turbulator Installment in a Shell Tube Heat Exchanger: A CFD-Based Simulation Approach by Using Nanofluids
Hybridized ant colony algorithm for the Multi Compartment Vehicle Routing Problem
Bulk wheat transportation and storage problem of public distribution system
References
Metaheuristics in combinatorial optimization: Overview and conceptual comparison
A Note on the Generation of Random Normal Deviates
Performance-effective and low-complexity task scheduling for heterogeneous computing
Related Papers (5)
A genetic algorithm for task scheduling on heterogeneous computing systems using multiple priority queues
Energy-Efficient Stochastic Task Scheduling on Heterogeneous Computing Systems
Frequently Asked Questions (12)
Q2. How do you generate a molecule's priority queue?
In order to achieve a good uniform coverage, the priority queues are generated by selecting from left to right an atom in the priority queues for these molecules.
Q3. What is the heuristic rank policy for the task scheduling problem?
In order to achieve a good “seeding” for the task scheduling problem, the authors take advantage of the heuristic rank policies [18], which are mostly used by traditional list scheduling approaches for estimating the priority of each subtask.
Q4. What is the reason why the makespan decreases with the increasing number of computing processors?
When the number of computing processors approaches the degree of parallelism, further increasing the number of computing processors will be of little help in reducing the makespan.
Q5. What is the size of the search space of the heuristic CRO method?
The size of search space of the heuristic CRO method proposed in this paper is much smaller than n!, and in order to find increasingly better solutions (i.e., the molecule structures with less energy), the operations simulating the four types of chemical reactions have to be performed over the solutions by using the heuristic transformational approach.
Q6. How many iterations are there in the search loop?
The algorithms are terminated when the value convergences to a relatively stable state (i.e., the makespan remains unchanged) for a preset number of consecutive iterations in the search loop (in the experiments, it is 10000).
Q7. What does the operator of inter-molecular ineffective collision do?
It means that the molecules ω′1 and ω ′ 2 can have a wide range of random search space than those of on-wall ineffective collision operator can do.
Q8. What is the DAG topology of an exemplar application model?
B. Application ModelIn this work, an application is represented by a DAG graph, with the graph vertexes representing tasks and edges between vertexes representing execution precedence between tasks [6].
Q9. What is the heuristic method used to find the execution order of the tasks?
A solution ω of the execution order is encoded as an integer queue and an integer represents a task id, i.e., ω = {T1, T2, · · · , Ti, · · · , Tn}.
Q10. What is the average performance of the proposed HCRO algorithm?
Figs. 25 and 26 show that the proposed HCRO algorithm outperforms HEFT-B and CPOP, and it can achieve a better average performance than DMSCRO with lower overhead.
Q11. How can the HCRO operator of decomposition escape from local optima?
The approach can effectively help the HCRO operator of decomposition to escape from local optima, by keeping the maximum hamming distance between old molecule ω and new molecule ω′1.
Q12. Why is the number of solutions in the CRO process varied?
Note that because of the decomposition (and the synthesis operator discussed later) operator, the number of solutions in the CRO process may be varied during the reactions, which is a feature of CRO that is different from genetic algorithms.