Optimal design of linear sensor networks for process plants: A multi-objective ant colony optimization approach

• Construct a multi-objective design model of non-redundant linear sensor networks.
• Develop an efficient multi-objective ant colony optimization method.
• Provide an integrated tool for multi-objective optimal design of sensor networks.

This investigation is performed to study nonredundant linear sensor network design problems that simultaneously optimize three objectives, namely cost, precision and reliability. In this article, a novel ant colony algorithm is proposed to efficiently tackle this multi-objective combinatorial problem. Both heuristic information and phenomenon are described as multiple matrices and a dynamic random weighted strategy is introduced to compute selection probability. A linear independent relationship is presented to guide a feasible solution construction process. An external archive is used to improve the convergence speed. Moreover, the TOPSIS method is adopted to aid in multi-criteria decision making with respect to Pareto-optimal solutions. The proposed method is successfully applied to a multi-objective design problem of a steam metering network of a methanol production plant. The results illustrate that the proposed method could find a set of global Pareto-optimal solutions, which is helpful for understanding the relationship among cost, precision and reliability, and consequently is helpful for final decision support. Therefore, the integrated methodology of a multi-objective ant colony algorithm and a multi-criteria decision making technique provides a promising tool for the optimal design of sensor networks by simultaneously considering multiple conflicting objectives.