Page 18 - Mathematical Models and Algorithms for Power System Optimization

8 Chapter 1

As for data processing and mining, the logic and rationality of the original input data need to be
considered, because if there is no logical relationship among the data, then the data is
meaningless like sand, even for huge amounts of data. This is a prerequisite to obtain a feasible
solution for the optimization calculation. For the original information processing, this book
proposes a method to confirm a solution space for linear programming to avoid no solution
(Chapter 2), proposes a mathematical conversion model between different interval
measurements to reduce the measurement data (Chapter 7), and demonstrates the equivalence
between local information and global control to ensure the effectiveness of local control
(Chapter 8).
As for the applications of AI technology in this book, the stochastic search algorithm is
applied as a nonnumeric and integer method, which is used to relax the hard boundaries of
variables and equations and form a better starting point. Since Simulated Annealing (SA)
method can be used to represent nonlinear, mixed integer problems, this book combines it with
the Newton-Raphson method to create new algorithms to solve nonlinear power flow problems
(Chapter 4). Since Genetic Algorithm (GA) is a random search algorithm with the ability to
obtain a global optimal solution, it is a nonnumeric and integer method, and can be used to
search for better integer solutions. This book combines the expert rules, fuzzy mathematical
concepts, and GA algorithms with traditional optimization methods to improve the possibility
of obtaining discrete solutions (Chapter 6). Since the expert rules can be used to create a
knowledge base, this book uses it to accumulate the experience of operational personnel to
improve the efficiency of a generator maintenance scheduling (GMS) solution (Chapter 3).