online-resource-scheduling

1# Online Resource Scheduling
2 
3Use this skill to build online schedulers that make deterministic decisions from the current observation only.
4 
5## Core Workflow
6 
7Convert each observation into a temporary state, rank pending work, score feasible actions by weighted marginal cost, update the temporary state immediately, then replay the final action list before returning it.
8 
9```text
10actions = []
11temporary_state = copy_resources(observation)
12 
13for item in ranked_pending_items(observation):
14  candidates = enumerate_feasible_actions(item, temporary_state)
15  if not candidates:
16    actions.append(defer_or_reject(item))
17    continue
18 
19  scored = []
20  for action in candidates:
21    deltas = estimate_objective_deltas(action, temporary_state)
22    score = sum(weights[k] * deltas[k] for k in deltas)
23    scored.append((score, stable_tie_break(action), action))
24 
25  chosen = min(scored)[-1]
26  actions.append(chosen)
27  apply(chosen, temporary_state)
28 
29validate(actions, observation)
30return actions
31```
32 
33## Weighted Marginal Scoring
34 
35When a task provides objective weights, use them to compare feasible actions. Avoid fixed rules such as "always first-fit", "always minimize fragmentation", or "always use the tightest slot". Those can be wrong when another objective component has a larger weighted effect.
36 
37Suggested generic workflow:
38 
391. Read visible objective weights.
402. For each pending item, enumerate feasible actions.
413. For each feasible action, estimate the change in each objective component.
42…