Cost modelling provides a unified framework for optimising industrial processes by simultaneously considering all the key metrics that characterise the performance of a process. The broad scope of cost modelling is to correlate the process parameters with the plant performance metrics and, in turn, relate these to the overall process cost. In the present paper, the concept of cost modelling will be elaborated in a case study on the gas carburising operation. The cost model for the case carburisation process relates the control-lable parameters (carbon potential and T-t set points) to the relevant cost drivers (energy, productivity, gas consumption, emissions) and quality parameters (case depth, grain size and distortion). These cost drivers and quality parameters are transformed to obtain the overall normalised production cost using appropriate cost functions. The cost model is subsequently used to optimise the carburisation cycle by the exhaustive search method for boost stage optimisation and by the differential evolution method, a genetic algorithm based technique, for the entire cycle optimisation, with the objective of minimising the production cost. This approach resulted in significant energy reduction (14%) and productivity enhancement (20%) in an industrial carburising operation. © 2004 Institute of Materials, Minerals and Mining.