Control charts are used to provide warnings in a process where the quality level is determined by more than one correlated quality characteristic. Recent studies have shown that the use of a variable sample size (VSS) design results in charts with higher statistical power in detecting and identifying small changes in the mean vector compared to a fixed sampling rate design. In this paper, a new control chart design using the Markov chain approach is designed, in which in addition to the variable sample size, the control limits are also variable. In this design, in addition to considering statistical criteria, the economic optimization of the design is also considered using the economic model of Costa and Rahim (2001). This economic model includes such things as the cost of false alarms, the cost of identifying deviations due to causes and repairing the process, the cost of producing the product when the process is out of control, and the cost of sampling and inspecting the products. Also, the genetic algorithm (GA) approach is used to optimize the desired model and obtain the optimal parameters of the model. Finally, the graphs of and are compared with the expected cost per unit of time.