SIH26027 · Ministry of Railways · Smart India Hackathon 2026
Three registers.
One corridor. One plan.
Engineering, S&T and Traction each keep their own maintenance register, and each asks the corridor for time separately. Nothing in the present process puts the three in front of a division together, so work that could travel in one block is granted as three.
Corridor plans them jointly. Asset risk and train detention are priced in one unit — detention-minute equivalents — so an overdue rail flaw and an overdue signal lamp can be weighed against the traffic each block would hold up. The plan then takes the work that removes the most risk for the least line occupation.
No department sees the other two
Engineering logs defects in TMS, S&T in SMMS, Traction in TDMS. The three do not meet in any system, so none can propose sharing a block even where it would obviously help. That is structural, not negligence.
Availability is not cost
A block at 07:30 and a block at 02:30 both appear free in the corridor. They are nothing alike in the detention they cause, and nothing in the current process prices the difference.
Priorities do not compare
Each register is worked most-overdue-first. But an overdue signal lamp and an overdue rail flaw are not comparable risks, and there is no shared unit to compare them in.
The measured result
Current practice versus the optimised plan
Every figure below is computed from the plan the engine produced — none of it is written in by hand. Both plans are scored by one function on one instance, with the same detention surface, the same risk prices and the same booked durations, and the baseline books the same P90 the optimiser books.
| Metric | Current practice | Optimised | Change |
|---|---|---|---|
| Blocks granted | 43 | 18 | -58.1% |
| Block-hours on line | 116.8 | 53.8 | -53.9% |
| Train detention (min) | 14,165 | 6,875 | -51.5% |
| Detention per task done | 329.4 | 143.2 | -56.5% |
| Residual risk carried | 8,549 | 4,919 | -42.5% |
| Tasks completed | 43 | 48 | +11.6% |
| Statutory closed | 21/35 | 30/35 | |
| Multi-department blocks | 0% | 72.2% | |
| Packing (work-hr / block-hr) | 0.82 | 2.06 | 2.51× |
Solver status
FEASIBLE
Deterministic — identical plan every run
Statutory obligation
30/35
Shortfall list names the binding constraint for each
Clubbed blocks
72.2%
13 of 18 blocks carry more than one department