The formula of two-part pricing for two groups of customers are drawn in this paper it isconcluded that the optimal marginal price depends on the differentiation of customers preference and themarginal cost.

Simulation shows that the optimal price is a decreasing function of initial inventory and an increasing function of sale period and potential buying probability;expected rev.

In this article,the optimal price model is built and solved by virtue of optimization in functional analysis.

A kind of smoothing factor,which optimizes general regression neural network (GRNN) by improved particle swarm optimization (PSO),is put forward and a method to forecast system marginal price by GRNN with optimized parameters is proposed.

The characteristics of system marginal price SMP are discussed according to economic theory.

To optimize the operation of power system,it is necessary to establish security constrained economic dispatch(SCED) and introduce locational marginal pricing(LMP)method.

On some fixed production price and comsumption level,there is a blind area of margrial price which the consumers and producers both agree on that if the price were made togher rightly,it would increase export surely while re.