Sensor location for linear networks is defined as a problem of multi-objective optimization.According to sensor precision

relative reconciliation precision is defined as the ratio of maximum reconciliation error to sensor measurement range.It is easy to compare relative reconciliation precision with sensor precision

thus it is more reasonable to determine optimization objectives and evaluate the design results.Two objectives are determined to minimize sensor location investment and to maximize relative precision of important variables.Some constraints are made to satisfy the network’s redundancy

control requirements

feasibility and some other conditions.Standard genetic algorithms are used to solve the optimization problem quickly and sufficiently.Fitness function is proposed by integrating virtual objective method and min-max method.Virtual objectives of investment and important variables’ relative precision are calculated by assuming that all streams are measured by the best sensors.A simulation result demonstrates its feasibility and shows that the proposed method accompanied by data reconciliation can significantly reduce sensor location investment.