In this study, the authors propose a modelling algorithm of inductors that provides a powerful tool for solving a variety of engineering optimisation problems. The algorithm has been developed in accordance with desire to consolidate cost benefit, losses minimisation and inductance maximisation, and finding the best manufacturing solution. A non-linear mathematical model for optimisation of three-phase, dry type double-E iron-core inductor has been presented in this study. Its purpose is to show that design of three-phase, air-cooled, network frequency inductors can be obtained using of new, adopted non-linear modelling algorithm-MISP (mixed-integer signomial programming), for minimising cost of materials (or inductor's mass) while ensuring the satisfaction of inductance value together with constraints imposed by overall measures, copper fill factor, maximum allowed induction, current density and temperature rise. The MISP is solved using branch-and-bound approach, specially implemented to inductor design problems. To confirm the result obtains by optimisation, inductor has been constructed according to the minimum loss criterion. The application is industry acceptable in cases if there are restrictions on the inductor weight. Also, where core dimensions are known, this approach can be used to design inductor by either maximising the inductance value or minimising total losses.