Table of contents
Binary particle swarm algorithm:
Code sharing for this article:
Article Summary:
An improved BS0 (binary particle swarm) method was proposed to solve the unit combination problem. First, the priority method is used to determine the initial unit combination. Based on this result, the range of the optimization window is determined, and BPSO is used to solve the problem within this range. During each iteration, heuristic adjustment strategies are used to ensure that the particles in each generation satisfy the constraints. Regarding the economic load distribution problem, the classic Lagrange multiplier method combined with the bisection method is used to solve the problem, which greatly improves the solution efficiency. Finally, the results obtained are compared with those obtained by other algorithms to prove that the proposed method has strong superiority and practicability.
Research Background:
The unit combination problem (UC) is one of the classic contents of power system optimization dispatch. Years of theoretical research and operation practice have shown that reasonable and optimized unit combination will greatly improve the economic benefits of power system operation. The unit combination problem in the power system is a large-scale mixed integer nonlinear optimization problem. The optimization goal is to rationally arrange power generation while meeting constraints such as system load demand, spinning reserve, and generator unit operation technical requirements. The start and stop of the unit and the output of the unit in each period minimize the power generation cost of the entire system.
In recent years, many scholars have done a lot of research work and proposed a variety of optimization methods for solving unit combination problems. Traditional methods include priority order method (PL), dynamic programming method (DP), mixed integer programming method (MP), Lagrangian relaxation method (LR), etc.; human T intelligence methods include genetic algorithm (GA), system evolution method (EP), simulated annealing (SA), particle swarm optimization (PSO), etc.
Compared with other methods, the priority method is fast in calculation and takes up less memory. This method was proposed earlier and is still being studied and applied. However, it often cannot find the optimal solution, and it is also difficult to deal with some dynamic constraints such as unit start and stop time constraints. The particle swarm algorithm has the characteristics of parallel processing and strong robustness. It can search for the global optimal solution with a greater probability, and its computational efficiency is higher than other traditional optimization methods. But if you search randomly in a larger range, it will also be blind to a certain extent. Both methods have certain advantages and disadvantages.
This paper combines the classic priority method with the binary particle swarm algorithm to solve the unit starting and stopping problem; the λ multiplier method is used combined with the bisection method for the load optimization problem. Through examples and comparative analysis of calculation results with other methods, it is verified that the proposed algorithm has good convergence and good results.
Binary particle swarm algorithm:
The late-stage convergence speed of the particle swarm algorithm is slow, the convergence accuracy is poor, and it is easy to fall into the local optimal solution. For this reason, many researchers combine the binary particle swarm algorithm and the chaos optimization algorithm, using the initial value sensitivity and ergodicity characteristics of chaotic variables to perform Particles that have lost their search ability perform chaotic search.
During an iterative optimization process of the particle group, at least one particle is in a stationary state, and other particles gradually approach the particle. When there is a particle that is close enough to the immobile particle, the particle can only search a limited area, and the optimization function is greatly weakened. In order to improve the search performance of this particle, this paper uses the chaos algorithm for initialization in the early stage of particle swarm optimization, and the initial particle swarm is selected.
The chaotic search method is used to separate particles that are about to overlap, and the distance between any particle and the current optimal particle is used as the criterion for whether the particles overlap. When the distance between the particles is less than the set value, the two particles are considered to have overlapped. At this time, the current optimal particle keeps its position unchanged, and the other particle is mapped to the chaotic variable space and performs chaotic motion based on the chaotic variable progression formula. Remap the obtained new chaotic variables into the variable search space to obtain new particles, and replace the original particles with the new particles obtained from the chaotic search.
Code running effect:
Code sharing for this article:
