this is for holding javascript data
Dmitry Volodin edited For_each_unit_we_define__.tex
about 8 years ago
Commit id: bf8648921ebe134ee3d85d725130613959c0f345
deletions | additions
diff --git a/For_each_unit_we_define__.tex b/For_each_unit_we_define__.tex
index c484840..2fe3c47 100644
--- a/For_each_unit_we_define__.tex
+++ b/For_each_unit_we_define__.tex
...
The Lagrangian is
\begin{multline}
L = \sum_u \hat{c}_u g_u + \sum_{j\in N} \lambda_j\left(\sum_{i: ij\in L}p_{ij} + \sum_{u\in j} g_u - \sum_{u \in j} d_u \right) \\
+ \sum_{(i,j)\in L} \left[\chi_{ij}( p_{ij} - B_{ij}(\theta_j - \theta_i)) + \sigma_{ij}(p_{ij} - p^{max}_{ij} \right] + \\
+ \sum_{i\in U} \left[\nu_u(g_u - \hat{g}_u) - \omega_u g_u \right]
\end{multline}
To simplify notation we will use $\lambda_u \stackrel{def}= \lambda_j, u\in j.$
The KKT conditions for basic system are as follows
\begin{align*}
\nabla_{g_u}L = -\hat{c_u} + \lambda_j + \nu_u - \omega_u &= 0, \quad \forall u\in U \\
\nabla_{p_{ij}}L = \lambda_j + \chi_{ij} + \sigma_{ij} &= 0, \quad \forall (i,j) \in L \\
\nabla_{\theta{j}}L =
\nu_u g_u &= 0 \\
\omega_u (\hat{g}_u-g_u) &= 0\\
\sigma_{ij}(p_{ij} - p_{ij}^{max}) &= 0 \\
\sigma, \nu, \omega & \geq 0
\end{align*}
For each unit we define the profit, having the LMP for the unit defined by
\lambda_u $\lambda_u =
\lambda_j \lambda_j$ we define the profit as
\begin{equation}
\pi_u = g_u(\lambda_u - c_u)
\end{equation}