TRANSMISSION // 2611.22892v1

Who gets a highway’s toll revenue when routes branch?

Source published

A theoretical study extends toll-sharing rules to networks with multiple routes. Its proposed method divides mileage and surcharge revenue separately, using route flows supplied as inputs rather than modeling drivers’ route choices.

AI-generated conceptual illustration for Who gets a highway’s toll revenue when routes branch?
AI-generated conceptual illustration; not documentary evidence.

When a pooled toll covers several road segments, how should the revenue be divided among them? Erfang Shan and Yan Chu develop a weighted rule for networks with branches and cycles. Their framework requires route-level flows or revenues as inputs; it does not predict which routes drivers choose or how they might change routes.

A settlement problem, not a toll-setting problem

A toll can cover several road segments, and a settlement rule determines how revenue from that toll is credited among those segments. This is different from setting tolls for drivers or dividing the cost of building a road. The paper models segments as the recipients of revenue; it does not assume a one-to-one match between segments and operators.

In Weighted Proportional Toll Sharing on General Highway Networks, Erfang Shan and Yan Chu extend Segments Equal Sharing (SES), a method developed for line highways, to networks with branches and cycles. Their proposed Weighted Proportional Sharing (WPS) method takes specified route flows—or route revenues—as inputs. It does not model how drivers select routes or reassign traffic after a change.

How the proposed rule works

The paper separates two parts of a toll: mileage revenue and fixed-surcharge revenue. For each route, WPS allocates mileage revenue among its segments in proportion to their weighted lengths. It allocates surcharge revenue among segments on that route with positive surcharges, in proportion to those surcharges. Segments with zero surcharge receive none of that component. The method adds contributions across routes, fully allocating each route’s revenue in each component.

For a given component, WPS matches route-by-route equal sharing when every positive-revenue route has a common, strictly positive weight across the segments sharing that component. For surcharges, eligibility matters: equal positive surcharge weights do not ensure equal-sharing equivalence if a route also includes zero-surcharge segments. The paper also defines route SES, which divides each route’s combined toll equally among all segments it uses.

The analysis is mathematical, not an empirical test of a real toll system. The authors define cooperative games in which road segments are players and route records determine the revenue available to groups. For mileage, a group receives revenue from routes whose full segment sets it contains. For surcharges, the model instead projects each route’s surcharge claim onto its positive-surcharge segments; those eligible segments can be disconnected if zero-surcharge segments lie between them.

The authors use these games to derive representations and characterizations. An axiom is a formal condition an allocation rule must satisfy. One set of conditions characterizes WPS using separability of mileage and surcharge allocations, route additivity, efficiency, treatment of unused segments, and proportional treatment of segments shared by every positive-revenue route. A separate characterization uses sub-highway efficiency and weighted average balanced contributions under a route-cancellation rule. Either set, on its own, uniquely identifies WPS within the model; the two sets are alternatives, not a combined requirement.

The paper also extends Segments Proportional Sharing, identifying it with the tau value of its route game, and characterizes a weighted, tree-based version of Segments Compensated Sharing.

Why route details matter

A network with cycles can offer more than one route between the same origin and destination. Origin–destination totals alone may not show which segments carried the traffic. Since WPS uses segment use on each route, different route splits can yield different segment allocations even when aggregate trip volumes and toll receipts match.

The authors demonstrate this with a constructed four-segment network and two scenarios. Traffic shifts between two routes while flow on the third route, from a to b via s, stays fixed. The scenarios preserve origin–destination volumes by vehicle class and aggregate mileage revenue, surcharge revenue, and total toll revenue. Yet segment allocations differ under both WPS and route SES. This illustrates how the rules respond to route inputs; it is not evidence about actual drivers or a real highway.

Applying the method would require route records, route shares supplied by an external traffic model, or clearly specified scenarios when several routes serve the same origin–destination pair. In the paper’s route-cancellation comparison, every specified route containing a selected segment is cancelled while other route records are retained. This is a defined counterfactual, not a prediction of what drivers would do after an actual closure. Modeling that response would require a separate traffic-reassignment model.

What the results do—and do not—show

The central findings are theoretical. The authors show that WPS is the sum of two weighted Shapley allocations—a game-theory approach that distributes coalition dividends in proportion to specified weights—and present two alternative axiomatic characterizations. These results establish how the rules behave under the paper’s assumptions. They do not show that WPS is fair in every policy context, would be adopted by operators, or improves traffic or welfare. The chosen weights, eligible segments, and route inputs all matter.

The source is an arXiv preprint. Its numerical example is a constructed illustration, not an evaluation of a toll system in operation. The paper also identifies negative route revenues and coalition-driven rerouting as cases needing separate analysis. Read the results as a framework for allocating revenue given specified routes and charges, not as a tested recommendation for a particular highway.

Automatically matched by shared topic keywords. These links do not imply that the sources agree or establish a consensus.