In this paper, we adapt co-evolutionary multi-task learning for multi-step prediction where predictive recurrence is developed to feature knowledge from previous states for future prediction horizon.
The goal of the paper is to present a network architecture with predictive recurrence which is capable of multi-step prediction through a form of multi-task learning.
Then, the fuzzy PID control based on the multi-step prediction is introduced to acquire the optimal control law.
Because the overall control law is a linear one, this design gives a direct and also effective multi-step prediction method and avoids the complicated nonlinear optimization.
However, small sample size and low precision of multi-step prediction limits its application in battery RUL estimation for sparse RVM algorithm.
The technique is based on the multi-step prediction error and on knowledge of the system initial condition and it takes an advantage of the well-known standard relations from the area of state estimation techniques and least square method.
The set of goals and objectives used in the MOGA model optimisation reflect different requirements in the design: obtaining good generalisation ability, good balance between one-step-ahead prediction accuracy and model complexity, and good multi-step prediction accuracy.
The set of goals and objectives used in the model optimisation reflect different requirements in the design: obtaining good generalisation ability, good balance between one-step-ahead prediction accuracy and model complexity, and good multi-step prediction accuracy.
A multi-step prediction procedure employing the Parallel, Levy and Effective Medium Theory models sequentially for the components other than ice and air, ice and then air respectively is recommended.
In this paper, the generalized predictive control (GPC) is used as one of existing controllers, because the GPC is designed based on multi-step prediction, and is effective for systems with large time-delays, ambiguous time-delays and/or time-variant time-delays.
Obeying this principle facilitates minimization of a parameter identification cost function which is the sum of long-range (multi-step) prediction error squares instead of the sum of single-step prediction error squares.
