Question regarding convergence of integer optimization

[yolking]

Hello! I am using pre-release version because I am interested in integer optimization. So far on the one hand I get pretty fast close to global optimum using it.

from bayes_opt import SequentialDomainReductionTransformer
pbounds = {
    "x1": (1, 1000, int),  
    "x2": (1, 1000, int),
    "x3": (1, 1000, int),
}
optimizer = BayesianOptimization(
    f=opt_func,
    pbounds=pbounds,
    verbose=2, 
)

best_target_ind = 0
best_target = -1000000000000
for i in count():
    next_point = optimizer.suggest()
    target = opt_func(**next_point)
    optimizer.register(params=next_point, target=target)
    if target>best_target:
        best_target = target
        best_target_ind = i
    if i-best_target_ind>150:      
        print(f"\t{next_point}")
        break
print("Best result:")
print(optimizer.max)

But BO seems to make very little effort to improve found optimum and continues to look far away from found optimal point. I checked docs for possible solutions to make it look closer to found optimum. SequentialDomainReductionTransformer may have been one of them, but it isn't available currently for integer optimization. Changing acquisition function after N unsuccessful iterations to
acquisition.ExpectedImprovement(xi=0.0)
seems like another possible solution, but it doesn't seem to do much.
Here is chunk of log of my optimization. The best optimum was found on 270 evaluation 0.5739394489614946 at [129, 2,740] and after that for 150 iterations no better point was found. Unreached global optimum is 0.5751567341679235 at [128, 2, 711]

270. 0.5739394489614946 : ( 129, 2, 740)
271. 0.562122923291349 : ( 141, 433, 755)
272. 0.5637633205086482 : ( 133, 591, 706)
273. 0.5671965653573532 : ( 137, 75, 682)
274. 0.5648119388511881 : ( 135, 384, 718)
275. 0.5556753869790383 : ( 168, 94, 804)
276. 0.5391487427464875 : ( 201, 750, 693)
277. 0.5620342532315347 : ( 135, 575, 757)
278. 0.5094327951926114 : ( 441, 6, 710)
279. 0.49312577749028136 : ( 325, 854, 978)
280. 0.5717251849734983 : ( 129, 45, 700)
281. 0.4993160246208407 : ( 389, 388, 934)
282. 0.5616862869471217 : ( 112, 19, 756)
283. 0.5668709566752652 : ( 141, 94, 728)
284. 0.5412575931496266 : ( 120, 235, 856)
285. 0.5260654968073734 : ( 293, 224, 772)
286. 0.5607839319758643 : ( 164, 1, 808)
287. 0.5570700592187176 : ( 154, 426, 727)
288. 0.5679625492297719 : ( 132, 2, 789)
289. 0.4135498252513754 : (1000, 646, 641)
290. 0.5578181264865069 : ( 145, 265, 795)
291. 0.4923969331624663 : ( 559, 1, 709)
292. 0.5713064661872675 : ( 132, 29, 701)
293. 0.5496693450599132 : ( 127, 287, 581)
294. 0.5608862509741278 : ( 134, 475, 781)
295. 0.5653586598894028 : ( 123, 395, 676)
296. 0.4899158653866104 : ( 399, 95, 999)
297. 0.5738924692628588 : ( 125, 1, 699)
298. 0.5649884129475095 : ( 150, 6, 754)
299. 0.5403166987731776 : ( 198, 545, 821)
300. 0.5048640345902715 : ( 481, 590, 863)
301. 0.5152179966179371 : ( 420, 8, 849)
302. 0.5695567422933128 : ( 122, 76, 700)
303. 0.5526048705799629 : ( 153, 687, 682)
304. 0.551958875136273 : ( 189, 3, 680)
305. 0.5220407644276102 : ( 329, 136, 877)
306. 0.5624521509865348 : ( 139, 479, 704)
307. 0.5624314035303094 : ( 133, 123, 790)
308. 0.5726957026706154 : ( 132, 3, 743)
309. 0.5574865183309355 : ( 147, 601, 733)
310. -0.003711854927608495: ( 991, 338, 14)
311. 0.5717020699778077 : 133, 22, 703
312. 0.5659367577295359 : 135, 5, 633
313. 0.5712659572231308 : 125, 29, 678
314. 0.5639185645465573 : 146, 2, 798
315. 0.5677979749663973 : 119, 70, 674
316. 0.5721542636991069 : 129, 32, 692
317. 0.5728171813402892 : 124, 19, 694
318. 0.5727085645868546 : 137, 2, 727
319. 0.5733057101998734 : 128, 20, 693
320. 0.5622654692128538 : 117, 186, 651
321. 0.5730132343477814 : 132, 7, 716
322. 0.5713460763264467 : 131, 23, 684
323. 0.5722638791551792 : 122, 2, 741
324. 0.4259971144997654 : 997, 265, 991
325. 0.5712120332543718 : 139, 6, 709
326. 0.5669250140427018 : 139, 91, 716
327. 0.4380931357144683 : 543, 1, 495
328. 0.5702820490066987 : 136, 39, 720
329. 0.5714631035555393 : 119, 1, 726
330. 0.48784666007730465 : 539, 827, 739
331. 0.5711345693363578 : 132, 27, 693
332. 0.5640471909903313 : 134, 229, 767
333. 0.5713302396633932 : 136, 20, 718
334. 0.5705349537264771 : 121, 29, 674
335. 0.5685407764156462 : 122, 10, 769
336. 0.5709470591129177 : 130, 31, 748
337. 0.5717805631344037 : 127, 37, 711
338. 0.5706775254772627 : 133, 1, 767
339. 0.5654006178672313 : 127, 161, 760
340. 0.5696959478895336 : 115, 2, 706
341. 0.56588772894393 : 121, 203, 728

Is this expected behavior? Other algorithms like DE and CMA become much more interested in convergence around optimal point after some time. Is it currently possible after, say, 100 unsuccessful iterations to look for optimum close to best point so far? Probably modifying bounds by hand or starting again with smaller bounds is one possible solution.

[till-m]

Hey @yolking,

would it be possible for you to post a complete snippet, including the function you're optimizing, so that I can run the code and check for myself?