By grid optimization, the determination of the size of the fed zone, when per unit cost of electricity delivery – the cost referred to transmitted energy – is minimum, is understood. Per unit cost is a function of substation cost and of fed zone grid indexes: length of the power lines and the amount of their material, losses of energy and of voltages which, in turn, are functions of the fed zone size. The indexes are different for various fed zone shapes. When outlines of a fed zone are not defined topographically, the most profitable shape of the zone is hexagon. However, the square or some other fed zone shape sometimes can be more acceptable due to the location of consumers. These indexes are defined in the paper. Besides, it turns out that, when minimizing the per unit cost of electricity delivery, a binomial of a fed zone main size is insufficient to represent the said per unit cost; it is shown here that a trinomial or even four-term polynomial is necessary to represent the per unit cost of electricity delivery