We bring methods and approaches to mathematically formulate the Inverse Problem into an Topology Optimization one. For instance, with a target magnetic induction or field B0, we want to design a device which will be able to produce this value at some fixed points. In this case, the objective is to minimize the gap between a computed value B and B0 at these points. For such a problem the variables are the device sizes, the material properties and the sources. Without lost of generality, we consider as variables the material properties, and the process is to fill the design domain with iron (value 1) and void (value 0); it is why this type of problem is called a 0-1 problem. The magnetic field B values are obtained by solving Maxwell's Equations via the Finite Element software FEMM (Finite Element Method Magnetics). This 0-1 problem is relaxed into the interval [0, 1], then the obtained continuous problem is solved with steepest descent algorithms. We compute it by using the Adjoint Variable Method of design Sensitivity Analysis for magnetic circuits. Numerical resolutions of our Topological Optimization is performed with a function of Matlab's Optimization Toolbox: fmincon. In general, the numerical results contain some intermediary values and these solutions are not manufacturable. Then, we avoid these intermediary values by using the SIMP method (Solid Isotropic Material with Penalization). This method combined with the relaxed problem give us binary solutions.