A Data-Characteristic-Driven Decomposition Ensemble Forecasting Research on the Demand of Space Science Payload Components
Algorithm 1
MPSO algorithm flow.
Input: time-series data, number of particles , number of parameters , number of iterations , inertia factor , individual
learning factor and social learning factor
Output: the position of the optimal parameter set
1 Set , randomly generate chaotic variable according to formula (12)
2 According to formula (13), the initial position is obtained, and the initial velocity is randomly generated
3 Calculate the fitness function according to formula (11)
4 Assign the individual historical optimum position directly, calculate the global historical optimum position , and make
5 While ()
6 Generate and randomly, adjust the inertia factor dynamically according to formula (26), and update the position and velocity of each particle according to formula (9) and formula (10).
7 Calculate the best value and the worst value of the population in this iteration, perform population crossover and mutation according to formula (14)–formula (25), and update the position of particles.
8 Calculate the fitness function according to formula (11) and update the global historical optimum position .
