Skip to content

Feedback from Bolin-Z/ChargingGame #270

@toruseo

Description

@toruseo

I found that @Bolin-Z did extensive application of UXsim (and UXsimpp) for their work and obtained many insights on UXsim (and UXsimpp) itself. Thank you very much!

Summary (curated by AI (Claude Code))

  • [Bug][High] A1 Ensure unique Vehicle naming in adddemand by adding a global counter or equivalent mechanism to prevent name collisions under same OD and timestep
    • this is on uxsimpp
  • [Bug][Low] A2 Correct misleading comment in route_search_all (currently states Floyd-Warshall, actual implementation uses Dijkstra)
    • this is on uxsimpp
  • [API][Medium] B1 Remove side effect in Analyzer.__init__ that automatically creates output directories; implement lazy directory creation upon result export
  • [API][Low] B2 Clarify semantics of Link.actual_travel_time(t) in documentation, or consider renaming to estimated_travel_time to reflect that it returns an updated estimate, not a future prediction
  • [API][Medium] B3-B4 Explicitly document time units (seconds vs timestep) for all time-related attributes, including log_t_link, arrival_time, departure_time, and departure_time_in_second
  • [Documentation][High] C1 Add warning that large deltan values may systematically truncate low-flow OD demand; optionally emit warning when residual cumulative flow remains below deltan at simulation end
  • [Documentation][Medium] C2 Clarify risks when using self-loop links (node -> node), including potential infinite reselection under DUO and ambiguity in link exit detection; recommend predefined routes or emit warnings on creation
  • [Documentation][Medium] C3 Explicitly explain the overwrite-based update strategy of traveltime_actual and its importance for UE-DTA convergence
  • [UE-DTA][High] D1-D6 Improve convergence mechanism and criteria, including gap-sensitive and time-decaying switching probability, completed ratio threshold, stability condition on recent GM variance, and immediate termination upon convergence detection
    • this is intentional
  • [Performance][Reference] E1 Address simulation core bottleneck (approximately 77 percent runtime share); consider low-level implementation (C, C++, Rust) for performance-critical components

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't workingenhancementNew feature or request

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions