In integrated steel works, byproduct gases are generated in the iron and steel making process, which accounts for approximately 30% of the total energy involved. The efficient utilisation of these gases is significant for energy saving and CO2 reduction in the iron and steel industry. In this paper, a mixed integer linear programming (MILP) model was proposed to optimise the byproduct gas management for the minimisation of operation costs and the efficient usage of energy. Compared with previous models, this proposed model considered the influence of the boiler penalty factor (BPF) and gasholder penalty factor (GPF) on optimisation results. The sum of the standard deviation volume (SSDV) and total switching times (TST) are defined to evaluate the effect of penalty factors on gasholder and boiler stability. The results of a case study indicate that the SSDV and TST are sensitive to the GPF and BPF, i.e., penalty factors have a large impact on optimisation results. Because the SSDV and TST are two confronted variables, Pareto optimality was applied to identify reasonable penalty factors which were used in the MILP model to obtain reasonable optimisation of the byproduct gas system. The optimisation results demonstrate that, compared with manual operation, the planning of the optimal distribution of byproduct gases proposed in this study can reduce the fluctuation of the volume of the gasholders and the load of the boilers to make the operation of the byproduct gas system safe and stable. Furthermore, according to sensitivity analysis, the stability of gasholders and boilers are sensitive to electricity price change. (C) 2015 Elsevier Ltd. All rights reserved.