It is becoming increasingly evident that in the future the Internet will host large numbers of software agents that aid or even act on behalf of companies and consumers to make decisions and carry out transactions. In this paper, we develop different pricing and ordering strategies with limited information requirements for e-retailers under the electronic business environment, and study the dynamics and performance of these strategies through multi-agent simulation. The results not only reproduce some phenomena in former literature, such as the "tacit collusion" and the cyclical price wars,but also generate new findings, for example, the "tacit collusion" does not always exist among hill-climbing e-retailers, and the cyclical price wars will be suppressed when e-retailers consider their competitors'strategies. The results show that in electronic business environments, to test different strategies and evaluate their performance by simulation before they are employed into practice is of great importance.