Industrial Wireless Sensor Networks (IWSNs), a novel technique in industry
control, can greatly reduce the cost of measurement and control and improve
productive efficiency. Different from Wireless Sensor Networks (WSNs) in
non-industrial applications, the communication reliability of IWSNs has to be
guaranteed as the real-time field data need to be transmitted to the control
system through IWSNs. Obviously, the network architecture has a significant
influence on the performance of IWSNs, and therefore this paper investigates
the optimal node placement problem of IWSNs to ensure the network reliability
and reduce the cost. To solve this problem, a node placement model of IWSNs
is developed and formulized in which the reliability, the setup cost, the
maintenance cost and the scalability of the system are taken into account.
Then an improved adaptive mutation probability binary particle swarm
optimization algorithm (AMPBPSO) is proposed for searching out the best
placement scheme. After the verification of the model and optimization
algorithm on the benchmark problem, the presented AMPBPSO and the
optimization model are used to solve various large-scale optimal sensor
placement problems. The experimental results show that AMPBPSO is effective
to tackle IWSNs node placement problems and outperforms discrete binary
Particle Swarm Optimization (DBPSO) and standard Genetic Algorithm (GA) in
terms of search accuracy and the convergence speed with the guaranteed
network reliability.