An Application of a Game of Discrete Generalised Pursuit Automata to Solve a Multi-Constraint Partitioning Problem

This paper presents a Learning Automaton (LA) solution to the Multi-Constrained Mapping problem, which has its applications in the allocation of processes on processors so as to satisfy multiple (possibly conflicting) constraints. Mathematically, it considers the problem of partitioning a set of elements (or objects) into mutually exclusive classes (or groups), where elements which are "similar" to each other are, hopefully, located in the same class. This problem has been shown to be NP-Hard, and the literature reports solutions in which the similarity constraint consists of a single index. For example, typical "similarity" conditions that have been used in the literature include those in which "similar" objects are accessed together (as in the context of query systems), or when they communicate (as processes do) with each other. The application at hand is the static mapping problem (SMP) of distributing the processes of a parallel application onto a set of computing nodes. Such an application may run on multiple GRID sites where it is desirable avoid centralised control and mapping. This paper proposes a solution to this combinatorial optimization problem resulting from the collective behaviour of independent Discrete General Pursuit Automata (DGPA) that tries to learn the digraph of the communication among the processes of the application, and group together processes with strong mutual dependencies. Earlier learning solutions to the problem were either based on centralised mapping with full system knowledge. In this paper, we attempt to relax this assumptions, thus rendering the problem more complex. The present solution performs very well when the system size is small. However, the simulated results demonstrate that the quality of the final solution decreases with the number of elements. Thus, although this is the first reported solution to the problem which incorporates the specific digraph properties of the objects, the scalability of the solution to the problem which incorporates the specific digraph properties of the objects, the scalability of the solution remains open.

[1]  M. A. L. THATHACHAR,et al.  A new approach to the design of reinforcement schemes for learning automata , 1985, IEEE Transactions on Systems, Man, and Cybernetics.

[2]  Francine Berman,et al.  On Mapping Parallel Algorithms into Parallel Architectures , 1987, J. Parallel Distributed Comput..

[3]  Olav Lysne,et al.  The Existence of a Network of Fixed-Sized Switches that Satisfies Any Communication Needs , 2004, PDPTA.

[4]  B. John Oommen,et al.  Generalized pursuit learning schemes: new families of continuous and discretized learning automata , 2002, IEEE Trans. Syst. Man Cybern. Part B.

[5]  C. Zheng,et al.  ; 0 ; , 1951 .

[6]  B. John Oommen,et al.  Stochastic learning-based weak estimation of multinomial random variables and its applications to pattern recognition in non-stationary environments , 2006, Pattern Recognit..

[7]  Kumpati S. Narendra,et al.  Learning automata - an introduction , 1989 .

[8]  F. A. Seiler,et al.  Numerical Recipes in C: The Art of Scientific Computing , 1989 .

[9]  El-Ghazali Talbi,et al.  Méthodes de placement statique des processus sur architectures parallèles , 1991 .

[10]  M. A. L. Thathachar,et al.  Networks of Learning Automata , 2004 .

[11]  Y.-K. Kwok,et al.  Static scheduling algorithms for allocating directed task graphs to multiprocessors , 1999, CSUR.

[12]  Peter G. Kropf,et al.  A Fast Distributed Mapping Algorithm , 1990, CONPAR.

[13]  Clement T. Yu,et al.  Improvements to an Algorithm for Equipartitioning , 1990, IEEE Trans. Computers.

[14]  Yoshio Mogami,et al.  A Consideration on the Learning Behaviors of the HSLA Under the Nonstationary Multiteacher Environment and Their Application to Simulation and Gaming , 2004, KES.

[15]  G. Horn,et al.  Towards a Learning Automata Solution to the Multi-Constraint Partitioning Problem , 2006, 2006 IEEE Conference on Cybernetics and Intelligent Systems.

[16]  Kaddour Najim,et al.  Learning automata and stochastic optimization , 1997 .

[17]  Shahid H. Bokhari,et al.  On the Mapping Problem , 1981, IEEE Transactions on Computers.

[18]  B. John Oommen,et al.  A fixed-structure learning automaton solution to the stochastic static mapping problem , 2005, 19th IEEE International Parallel and Distributed Processing Symposium.

[19]  B. John Oommen,et al.  Deterministic Learning Automata Solutions to the Equipartitioning Problem , 1988, IEEE Trans. Computers.

[20]  Mohammad S. Obaidat,et al.  Guest editorial learning automata: theory, paradigms, and applications , 2002, IEEE Trans. Syst. Man Cybern. Part B.