A continuous state transition algorithm with guaranteed optimality

As a constructivism-based intelligent optimization method, the state transition algorithm (STA) has exhibited powerful search ability in optimization. However, in basic STA, it usually sets the maximum number of iterations or function evaluations as the termination condition by experience. If the termination condition is not set well, it may stop at an unpredictable point. In this study, a continuous STA with guaranteed optimality is proposed by altering the termination conditions of the basic STA, without using the maximum number of iterations or function evaluations. It is proved that the modified STA can find a final solution with guaranteed optimality. Experimental results have testified the findings.