Mixed-integer quadratically constrained programming with application to distribution networks reconfiguration

This paper implements a mixed-integer quadratically constrained convex optimization program on distribution networks that have a radial structure topology for denigration of system's real power loss. The adopted mathematical model has been solved using Cplex optimization software. The global optimality of the obtained solution is guaranteed via convex relaxation of the continuous decision variables. This overcomes a major disadvantage of using artificial intelligence techniques, specifically meta-heuristic ones, that give no assurance on the solution quality or the optimality gap. Moreover, exact representation of the network losses through the assessment of the objective's function numerical-value is obtained. This overcomes the drawbacks of using linear approximations to power flow equations. The proposed algorithm has been evaluated on two IEEE standard systems and a realistic-sized network with a comparative study with other existing algorithms in the literature.