Pareto optimal solutions visualization techniques for multiobjective design and upgrade of instrumentation networks

Instrumentation networks are essential for the proper operation of chemical plants. This paper addresses the problem of designing and upgrading instrumentation networks. A multiobjective approach is selected to solve the optimization problem, instead of the traditional minimum cost formulation. This paper proposes two methodologies for the visualization of Pareto optimal solutions (POSs) for the design and upgrade of sensor networks: one is based on projections of the POS onto specific two-dimensional surfaces, and the second is the representation of the problem in parallel coordinates systems. The proposed network design algorithms enable the decision maker to see several candidate solutions for the network configuration, and at the same time, they allow the number of candidates to be narrowed to satisfy the decision maker's specifications.