A Hybrid Autonomic Computing-Based Approach to Distributed Constraint Satisfaction Problems

Abstract

Distributed constraint satisfaction problems (DisCSPs) are among the widely endeavored problems using agent-based simulation. Fernandez et al. formulated sensor and mobile tracking problem as a DisCSP, known as SensorDCSP In this paper, we adopt a customized ERE (environment, reactive rules and entities) algorithm for the SensorDCSP, which is otherwise proven as a computationally intractable problem. An amalgamation of the autonomy-oriented computing (AOC)-based algorithm (ERE) and genetic algorithm (GA) provides an early solution of the modeled DisCSP. Incorporation of GA into ERE facilitates auto-tuning of the simulation parameters, thereby leading to an early solution of constraint satisfaction. This study further contributes towards a model, built up in the NetLogo simulation environment, to infer the efficacy of the proposed approach.