WorkloadLeverScenario
Economics scenarios · Implementation: mlsysim/engine/scenario_definitions/economics.py.
Construct this scenario with keyword inputs. The fields below show the types and defaults declared by this class.
Attempt cost by term for several workloads, and three levers on cost per accepted task.
Each workload gives acceptance, environment, assessment in dollars, and either human (fixed per attempt) or handoff (charged per rejected attempt). The model term is model dollars, or the bill’s cached trajectory cost, or token_multiple chats of chat_input/chat_output tokens at the stated tariff.
Notes
Equation: The result property delegates to the named solver documented below; it converts declared scenario units before passing the inputs.
Assumptions: Input workloads declare acceptance and additive USD terms; cached bills and token multiples expand under one tariff.
Excludes: Interactions between workload levers or empirical prediction of acceptance improvement.
References
Derivation: the scenario composes mlsysim.solvers.WorkloadLeverModel. Their contracts supply the underlying equations; property equations show the remaining unit conversions and accounting.
Examples
>>> from mlsysim import Scenarios
>>> scenario = Scenarios.LighthouseWorkloadLevers
>>> value = scenario.result
>>> type(value).__name__
'WorkloadLeverResult'Attributes
| Name | Description |
|---|---|
| acceptance_gain | |
| input_price | |
| output_price | |
| result | Return the typed accounting result from WorkloadLeverModel using this scenario’s |
| workloads |
Methods
| Name | Description |
|---|---|
| terms | Normalize workload billing into USD-per-attempt terms, expanding cached bills, token |
terms
engine.scenario_types.WorkloadLeverScenario.terms()Normalize workload billing into USD-per-attempt terms, expanding cached bills, token multiples, and per-turn environment costs.
Equation: out.
workloads
Type: Dict[str, Dict[str, Any]].
Declared default: required.
input_price
Type: Quantity.
Declared default: required.
output_price
Type: Quantity.
Declared default: required.
acceptance_gain
Type: float.
Declared default: required.
result
engine.scenario_types.WorkloadLeverScenario.result
Return the typed accounting result from WorkloadLeverModel using this scenario’s declared inputs; units and exclusions follow that solver contract.
Equation: WorkloadLeverModel().solve(terms(), acceptance_gain).