The chapter concludes that the Particle Swarm Optimization based Lagrangian Relaxation approach provides the optimal solution for the Profit Based Unit Commitment problem by achieving low production costs, high profits, and minimizing execution time when parameters are properly selected. Future work could involve developing hybrid intelligent techniques to determine both unit commitment status and economic dispatch more quickly for large power systems, as well as extending the proposed approach to problems like Automatic Generation Control.
The chapter concludes that the Particle Swarm Optimization based Lagrangian Relaxation approach provides the optimal solution for the Profit Based Unit Commitment problem by achieving low production costs, high profits, and minimizing execution time when parameters are properly selected. Future work could involve developing hybrid intelligent techniques to determine both unit commitment status and economic dispatch more quickly for large power systems, as well as extending the proposed approach to problems like Automatic Generation Control.
The chapter concludes that the Particle Swarm Optimization based Lagrangian Relaxation approach provides the optimal solution for the Profit Based Unit Commitment problem by achieving low production costs, high profits, and minimizing execution time when parameters are properly selected. Future work could involve developing hybrid intelligent techniques to determine both unit commitment status and economic dispatch more quickly for large power systems, as well as extending the proposed approach to problems like Automatic Generation Control.
The foremost objective of this research is to formulate effective
solution techniques to optimize the power generation expenses in terms of reducing the total production cost and increasing profit. In the proposed four different techniques namely Modified Dynamic Programming algorithm, Dynamic Programming based Lagrangian Relaxation algorithm, Genetic Algorithm based Lagrangian Relaxation algorithm and Particle Swarm Optimization based Lagrangian Relaxation algorithm have been considered for obtaining the optimal solution for Profit Based Unit Commitment problem. In DPLR, GALR and PSOLR the economic schedule is obtained from Lagrangian Relaxation method. Equality constraints are satisfied in all the four approaches considering reliability as a main factor in solving profit based unit commitment problem. The reserve is added as a percentage of load to the generation while satisfying the equality constraints. The 10 unit system considered for evaluation is treated as an individual GENCO.
Modified Dynamic Programming approach employs a three stage
optimization algorithm for solving PBUC. In MDP the parameter lambda is optimized during the initial stage using the coordination equations. Lambda optimization is a repeated process till the condition is satisfied and therefore the convergence speed is slow. However the solution obtained is accurate than the individual DP solution. Constraint relaxation is allowed if the combinations do not converge at all during any stages of the scheduled time 133
period. The performance of the algorithm is evaluated for NTPS 7 unit and IEEE 10 unit system respectively.
Dynamic Programming based Lagrangian Relaxation avoids the
complexity experienced in MDP method. Lambda iteration method is employed in DPLR to obtain the unit commitment status. Forward Dynamic Programming is applied in solving the problem and while back tracking for optimal commitment it is ensured that the constraints are met. No constraint relaxation is observed in DPLR method. As the number of generating units is less the dimensionality issue is less concerned. UC status is determined by the DP method and the economic schedule is obtained by the Lagrangian Relaxation. The profit and total fuel cost of DPLR are not superior to MDP, but the convergence of the solution is fast. The performance of the algorithm is evaluated for 10 unit system.
Genetic Algorithm based Lagrangian Relaxation approach is based
on selection of the best population in the global search space. Selection of the optimal solution is based on the fitness function. In the proposed work the crossover probability percentage is higher than the mutation probability. This may avoid the solution to get struck in local optima. The advantage of GALR is the capability of finding out optimal UC schedules in less computing time, able to generate many different UC schedules and able to add more constraints for better optimum schedule. The results reveal that the production cost is less when compared to MDP and DPLR approaches interms of fuel cost, profit and convergence speed. Sometimes the determination of population size and mutation rate is time consuming which forces the algorithm to converge in a local maximum or minimum.
Particle Swarm based Lagrangian Relaxation approach eliminates
the drawbacks of all the other proposed approaches. The PSOLR requires less parameter tuning. There is less probability for the PSO algorithm to get struck 134
in local optima as it does not have overlapping and mutation calculation.
Performance of the algorithm is evaluated from their internal operators and parameter values. Proper selection of parameters in the PSO algorithm can further improve the quality of the solution.
The conventional methods solve PBUC problem by considering
only profit maximization and the demand constraint is not considered. But in the proposed methods the profit maximization and demand constraints are considered. Hence the proposed methods are the best for solving the PBUC problem.
To conclude, when the parameter values are properly selected,
PSOLR gives an optimal solution in terms of low production cost and high profit and consumes minimum execution time.
8.2 FUTURE WORK
Future work may be the application of hybrid intelligent techniques
for obtaining both the UC status and Economic Dispatch thus reducing the execution time for large sized practical power systems. The proposed technique can be extended to power scheduling problems of Automatic Generation Control. However, the research on the PSO is still at the beginning. Hybridizing PSO with the other intelligent optimization algorithms combines the advantages of the PSO with the advantages of the other intelligent optimization algorithms to create an even more better solution with wide application ranges.
In the simplest case, an optimization problem consists of maximizing or minimizing a real function by systematically choosing input values from within an allowed set and computing the value of the function. The generalization of optimization theory and techniques to other formulations comprises a large area of applied mathematics. More generally, optimization includes finding "best available" values of some objective function given a defined domain (or a set of constraints), including a variety of different types of objective functions and different types of domains.