standard tier · 72 briefs · 1,368 runs · results db2e24e

CityFormGen

Generative and optimisation methods for middle housing, compared under matched compute

RESEARCH GENERATIVE-DESIGN BENCHMARK - NOT CONSTRUCTION DOCUMENTATION OR CODE CERTIFICATION.

easy · maisonette · 2 storeys

B0058: a rowhouse on a 29.5 × 30.6 m lot

4 dwellings, 36 rooms. The pooled front across all methods holds 587 nondominated designs.

Requirements

Lot
29.5 × 30.6 m (903 m²)
Street side faces
120° (0 = north)
Setbacks front / rear / left / right
5.2 / 6.3 / 2 / 2.1 m
Storeys × height
2 × 2.8 m
Shared core
no, street-entered dwellings
Open space, net of parking
30% of the lot
Parking spaces
0
Courtyard
not required
Tightness (programme / buildable volume)
0.260

Programme

DwellingTypeStoreyMin m²Rooms
U02BRall63entry, living, kitchen, bathroom, circulation, bedroom primary, bedroom secondary, bathroom, circulation
U12BRall63entry, living, kitchen, bathroom, circulation, bedroom primary, bedroom secondary, bathroom, circulation
U22BRall63entry, living, kitchen, bathroom, circulation, bedroom primary, bedroom secondary, bathroom, circulation
U32BRall63entry, living, kitchen, bathroom, circulation, bedroom primary, bedroom secondary, bathroom, circulation

Required contacts: entry–living, kitchen–living. Doors forbidden between: entry–bedroom_primary, entry–bedroom_secondary, kitchen–bathroom.

Results by method

MethodSolvedHVEvalssNotes
Heuristic100%0.2638674.5
Random search0%0.00015006.8
Simulated annealing33%0.06915006.7
NSGA-II100%0.19215009.2
CP-SAT0%0.00004.4solver timeout
Random sketch + repair33%0.07715008.2repair collapse
Heuristic-seeded NSGA-II100%0.20315009.0

Mean over 1 seeds (the heuristic is deterministic: one run).

Representative designs from the pooled front

Each is best on one objective, or the knee: the most balanced trade-off. None is best overall.

Closest attempts that failed

For methods that found nothing feasible: their least-infeasible design, with its violations marked on its page.

Pareto explorer

Loading designs…

Convergence on this brief

Archive hypervolume against evaluations, mean over seeds.

Heuristic
00.261500 ev
Random search
00.261500 ev
Simulated annealing
00.261500 ev
NSGA-II
00.261500 ev
CP-SAT
00.261500 ev
Random sketch + repair
00.261500 ev
Heuristic-seeded NSGA-II
00.261500 ev

What stopped designs here

Share of each method's evaluated designs that violated each constraint.
methodC04C05C06C07C08C09C10C11
Heuristic7%53%41%0%6%1%1%
Random search14%94%99%1%90%99%39%
Simulated annealing18%72%81%1%77%61%22%
NSGA-II13%59%72%0%1%51%41%17%
Random sketch + repair23%45%62%2%5%87%85%24%
Heuristic-seeded NSGA-II11%53%57%0%2%20%21%9%

Constraint definitions