Annealing optimizer

The optimizer is used to optimize city blocks of certain land uses by building services of predefined capacity and area.

Prepare city

The method is time consuming and aimed to work with smaller territories, so we make a smaller model.

from blocksnet import City, AnnealingOptimizer

city = City.from_pickle('../model.pickle')

blocks = city.get_blocks_gdf(True)
nearest_blocks = blocks.sjoin_nearest(blocks[blocks.index == 1445], distance_col='distance')
nearest_blocks = nearest_blocks[nearest_blocks['distance']<=1000]
nearest_blocks.plot().set_axis_off()

city._blocks = {i:city[i] for i in nearest_blocks.index}

Method

import random
from blocksnet import LandUse
from blocksnet.method.annealing_optimizer import LU_FSIS, LU_GSIS

ao = AnnealingOptimizer(city_model=city)
blocks_lu = {block.id : random.choice(list(LandUse)) for block in city.blocks}
blocks_fsi = {b_id : LU_FSIS[lu][0] for b_id, lu in blocks_lu.items()}
blocks_gsi = {b_id : LU_GSIS[lu][0] for b_id, lu in blocks_lu.items()}
service_types = {st.name : 1/len(city.service_types) for st in city.service_types}
X, indicators, value, provisions = ao.calculate(blocks_lu, blocks_fsi, blocks_gsi, service_types, rate=0.95, max_iter=10_000)

Value : 0.464:  11%|█         | 1065/10000 [02:46<23:15,  6.40it/s]

print('Best objective value:', value)

Best objective value: 0.46475854673870937

{st : round(value,2)  for st,value in provisions.items()}

{'school': 0.68,
 'kindergarten': 0.22,
 'hospital': 1.0,
 'polyclinic': 0.34,
 'pitch': 0.0,
 'swimming_pool': 0.95,
 'stadium': 1.0,
 'theatre': 1.0,
 'museum': 1.0,
 'cinema': 1.0,
 'mall': 0.88,
 'convenience': 0.18,
 'supermarket': 0.5,
 'cemetery': 0.0,
 'religion': 1.0,
 'market': 0.84,
 'bowling_alley': 1.0,
 'university': 1.0,
 'playground': 0.09,
 'pharmacy': 0.79,
 'fuel': 0.5,
 'beach': 0.0,
 'train_building': 1.0,
 'bank': 0.51,
 'lawyer': 0.88,
 'cafe': 0.34,
 'subway_entrance': 0.45,
 'multifunctional_center': 0.03,
 'hairdresser': 0.47,
 'restaurant': 0.69,
 'bar': 0.46,
 'park': 0.0,
 'government': 1.0,
 'recruitment': 1.0,
 'hotel': 1.0,
 'zoo': 0.0,
 'circus': 0.0,
 'post': 0.13,
 'police': 0.0,
 'dog_park': 0.65,
 'hostel': 0.34,
 'bakery': 0.5,
 'parking': 0.02,
 'guest_house': 0.0,
 'reserve': 0.0,
 'sanatorium': 1.0,
 'embankment': 0.0,
 'machine-building_plant': 1.0,
 'brewery': 0.29,
 'woodworking_plant': 1.0,
 'oil_refinery': 0.0,
 'plant_of_building_materials': 0.79,
 'wastewater_plant': 0.0,
 'water_works': 0.0,
 'substation': 0,
 'train_station': 0.24,
 'bus_station': 0.26,
 'bus_stop': 0.07,
 'pier': 1.0,
 'animal_shelter': 0.07,
 'military_kom': 1.0,
 'prison': 0.0,
 'landfill': 0.0,
 'plant_nursery': 0.0,
 'greenhouse_complex': 0.0,
 'warehouse': 0.5}

ao.to_bricks_df(X).head()

	block_id	service_type	is_integrated	area	capacity	count
4	39	polyclinic	False	1400.0	150	1
28	39	parking	False	4750.0	190	1
58	39	market	False	6000.0	8000	1
82	39	school	False	4000.0	300	1
96	39	museum	True	1300.0	1000	1

ao.to_gdf(X, indicators).head()

	population	mall	plant_of_building_materials	polyclinic	playground	post	pier	landfill	kindergarten	animal_shelter	...	greenhouse_complex	reserve	bar	oil_refinery	bakery	multifunctional_center	embankment	hostel	geometry	land_use
block_id
39	1472	0.0	0.0	150.0	0.0	0.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	POLYGON ((352493.392 6644174.577, 352480.316 6...	residential
40	0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	POLYGON ((352846.893 6643739.658, 352854.917 6...	transport
41	0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	POLYGON ((352713.398 6643353.609, 352656.125 6...	business
42	0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	POLYGON ((352661.968 6643879.720, 352637.303 6...	transport
43	0	0.0	0.0	0.0	0.0	800.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	POLYGON ((353921.529 6644728.862, 353829.106 6...	recreation

5 rows × 69 columns