- identifying modes of behavior using sensitivity analysis - baseline and other sensitivity (can input prior draws from the following) - avoid serial table function (convolution), set model in equilibrium at normal point - parameterize table functions for sensitivity testing - can normalizing the input and output make the s-shape more symmetric? - allowing to track other stocks (e.g. obesity weight) ppl and their wiegith , workers and their experience, factories energy efficiency, tasks and error - Take out characteristic from its stock with the “average” density (exceptions need careful treatment, e.g. Rahmandad and Hu (2010)) ![[System Dynamics Review - 2010 - Rahmandad - Modeling the rework cycle capturing multiple defects per task.pdf]] Efforts to make lookup more analytic is parallel to (symmetric latent_data and parameter) - difference of sensitivity analysis of `SWEffxyz` and `xyz`? ![[Pasted image 20220930111546.png]] - ![[Pasted image 20220930111510.png]] 1 consumer and 2 producer, - T as total demand, C1 and C2 producer capa, P1, P2 is price, Q1, Q2 i) c1 + c2 >< T ii) p1 >< p2 | - | p1 > p2 | p2 > p1 | | --- | ------- | ------- | | T> | | | | T< | | | Allocation mechanism: robust m to n formulation (solve when demand and supply ) - energy going in and out five to five allocation. - allocation is based on subscript (allocation by priority) - capacity is not regarded so nested if then else becomes messy ![[Pasted image 20220930162931.png]]