Optimising Work Zones With Enhanced Cellular Automata and Delayed Mandatory Lane-Changing Modelling

In highway work zones, lane-changing behaviour near bottlenecks is more urgent than in general segments, referred to as delayed mandatory lane-changing (DMLC). This study proposes an enhanced cellular work zone automata model (DMLC-EWCA) that incorporates DMLC position probability distribution for micro-simulation. Analysing 12 cases from China, the study shows that lane-changing behaviour varies by lane and is influenced by traffic flow and work zone configuration. Extreme value theory is applied to model DMLC position distribution, and lateral lane-changing rules in DMLC-EWCA are improved. The results highlight that traffic flow, speed, and transition area length are key factors influencing vehicle DMLC position. The DMLC-EWCA outperforms the previous CA model at both microscopic and macroscopic levels. A case study applying the DMLC-EWCA optimised transition area length, enhancing innermost lane usage and overall traffic efficiency. The DMLC-EWCA offers a more accurate approach for modelling traffic flow in work zones.