Every number on this page is computed by the Chapter 9 notebook's own code. The sensitivities, the 10,000 situations of step 3 and the random constraint families of step 5 are fresh draws with the lab's random seed 26109, not Meridian records. Your notes stay on this page until you leave it; copy the memo before you go.
Meridian is preparing a reorganization and a board review of its limits. The chief operating officer wants to know which units really depend on which; treasury wants the reaction window for a demand shock; and the risk committee wants to know which reviews must be held jointly. You support the transformation office.
Use AXIOM to build Meridian's state architecture, find its feedback core and layers, put floors under shock arrival, test the summary maps, and design the joint-review calendar that the constraints require.
Worked Example 9.1; Tables 9.1–9.3; Figure 9.1. An arrow u → v asserts that v's next state reads u's state; a missing arrow asserts that it does not.
In the architecture builder, inspect the seven components, the fourteen typed arrows, and the constraints with their scopes; for each arrow, name the reading it carries and check that its source exposes that reading.
Run the topology analyzer: read off the strongly connected components, the condensation, and the layers, and compare them with the org chart.
Tarjan's algorithm finds the strongly connected components; a component with more than one member is a feedback core. Collapsing each to one node gives the condensation, which has no loops (Proposition 9.3); its levels are the causal layers. Shaded bands: drivers, feeder, core.
In the wavefront simulator, shock the market and watch which components it can have reached after one, two, and three quarters; weaken a dependence and confirm that arrival can come later but never earlier. Count the routes from the market to treasury with the path counter.
Breadth-first search from the market gives each component's distance, the earliest quarter-end at which a demand shock can change it (Theorem 9.2). Every declared arrow carries a positive sensitivity drawn with the lab's seed; switching arrows off is applied to that model.
10,000 situations drawn with the lab's seed, each arrow switched off with probability 0.3: the share in which each component first moves at its floor, later, or not within twelve quarters.
Routes (walks, which may revisit a component) from the market to treasury, by length: the entries of the adjacency matrix's powers (Section 9.8). Right: the five distinct paths.
In the aggregation inspector, build the two-box map, find the connections it invents, and drill down to confirm it; then aggregate by feedback cores instead and check that no connection is invented.
The matrix drills down: each number is the quarters a shock needs to travel from the row's component to the column's, a blank means no path, and ✕ marks a pair the summary map shows as connected and the wiring does not (Theorem 9.3).
On the consistency bench, enter the three planning rules of Theorem 9.4(i), confirm that each passes a bilateral check, and watch the joint check fail; then load Meridian's constraints, count the cycles, and run the validation checklist row by row.
Dates under rules “xj ≤ xi + b” are feasible exactly when no cycle of rules has negative total weight (Online Appendix 9.B.4).
Each coupling constraint links every pair in its scope (Definition 9.7). Right: group strategic risk on Chapter 1's digital plan against the ceiling of 75 and the risk committee's ±6 measurement error (Table 5.2).
No component moves before its distance from the shock (Theorem 9.2); summary maps add connections but never remove them (Theorem 9.3); and the joint-review burden is set by the number of cycles, not by the number of constraints (Theorem 9.4).
The declared map, the feedback core and layers, the arrival floors for a demand shock, the joint-review calendar, and the one missing arrow to test first.