Abstract: During the trading process of closed-end fund, some investors who have large amount of money may control the price of fund. On the basis of some given assumption, considering the condition that the "banker" investor for closed-end fund holding limited amount of initial investment at the beginning of a certain trade period, the fund price control process are divided into two stages. For the first stage, a single goal model which is transformed from a two goal model is gained with the known expected revenue ratio of every time of fund trade period, it can be solved by the Kuhn-Tucker condition. In the second stage, a model is constructed with minimum cash payment. It is a nonlinear program problem and the optimal price control order with an improved genetic algorithm is obtained. The calculating results of the simulation conforms with the "banker" investors' high selling and low buying process. It is valuable for the financial department to strengthen the supervision to the closed-end fund market, and it effectively prevents the excessively manipulation of fund price behavior for the closed-end funds trading.

Key words: closed-end fund, price, Kuhn-Tucker condition, nonlinear program, genetic algorithm