In this project, we study the use of learning automata to solve the nonlinear fractional bin
packing problem and its application to distributed web crawling.
Recent studies has been performed to solve the the so called nonlinear fractional knapsack
problem. The results were really fascinating and demonstrate the strength of learning automata
when applied to centralized resource allocation problems. Due to the ever growing deployment
of distributed systems, it becomes imperative to extend the later studies to cope with a
distributed context. The obvious and direct extension of the non linear fractional knapsack
problem is the non linear fractional bin packing problem. In this paper, we present two novel
approaches for solving the non linear bin packing problem. We relate this problem to resource
allocation problems and we show that it has applications within the field of distributed crawling .
An appropriate mapping of the nonlinear fractional bin packing problem to distributed polling
frequency determination problem is considered in this paper.
Comprehensive experimental results have showed the superiority of our proposed schemes
when applied to distributed web crawling.