Formation of policies guided by multivariable control theory

The literature associated with system modeling and decision-making falls short in automating policy formation for correlated multiple-policy multiple-objective (MPMO) processes. Enthused by the effectiveness of multivariable control theory, this paper proposes the implementation of a multiple-input multiple-output (MIMO) controller as a guidance tool for designing policies. System dynamics is considered for modeling the dynamic behavior of MPMO systems. The model is then converted to a state-space system where objectives and policies are mapped to time-varying reference trajectories and inputs, respectively. Subsequently, the design of policies driving the system outputs to meet certain objective profiles is converted to a multivariable control problem. This paper also recommends a class of multivariable controllers that is suitable for this domain of applications. Numerical simulations are included to illustrate the effectiveness of the proposed systematic approach.