Model deep dive

Every worksheet of the uploaded AI + Quantum pricing workbook, rebuilt here as a live running model. Each section below matches a sheet, but the numbers are computed by the engine rather than frozen in a cell.

00 · Overview 01 · Parameters 02 · Parsed lines 03 · Product model 04 · Price candidates 05 · Optimization result 06 · QUBO model 07 · Benchmark 08 · SHAP global 09 · SHAP local 10 · LIME local 11 · 30-minute engine 12 · Dashboard

Products

after name normalisation

Binary variables

93 feasible after the margin floor

Parsed order lines

769

1179 units sold

Optimality gap

0.0000%

quantum-inspired vs exact

00 · What this model does

Purpose, mathematics and honest limits.

The engine picks one price per product, all products at once, to maximise a blended business score: profit, stock sell-through and customer value up; expected spoilage and price movement down. It re-optimises every 30 minutes on refreshed demand and stock inputs.

The source file holds 400 orders and 769 parsed lines. It contains no cost, stock or spoilage data, so those are clearly-labelled synthetic assumptions you can edit on the Product model page.

The problem is written as a QUBO and solved twice — exactly, and with quantum-inspired annealing — so every run carries proof of how close the quantum-style answer is. No physical quantum hardware is involved.

Binary decision

x(i,k) = 1 when product i is sold at candidate price k

One price per product

Σₖ x(i,k) = 1 for every product i

Minimum margin

(Price − Cost) / Price ≥ 15%; other candidates disabled

Objective

Max Σ x(i,k) × [0.45·Profit + 0.20·Sell-through + 0.15·Customer value + 0.10·Waste score − 0.10·Volatility] × 100

Portfolio stability

0.98 ≤ average selected price index ≤ 1.02

QUBO Hamiltonian

H(x) = −Σ U(i,k)·x(i,k) + A·Σᵢ(Σₖ x(i,k) − 1)² + B·(AvgIndex − 1)²

01 · Parameters

Live engine defaults — change them on the Optimizer page and re-run.

Parameter	Value	Type	Notes
Minimum margin	15%	Constraint	Not in the source file — editable default.
Average price index min	0.98	Constraint	Portfolio stability lower bound.
Average price index max	1.02	Constraint	Portfolio stability upper bound.
Target price index	1.00	Objective	Quadratic pull toward today's average price.
Re-optimisation cadence	30 min	Runtime	Scheduled engine refresh.
Price candidates	0.90×, 0.95×, 1.00×, 1.05×, 1.10×	Search space	Rounded to the nearest ₹5.
Profit weight	0.45	Objective	Largest single driver.
Inventory weight	0.20	Objective	Rewards sell-through.
Customer value weight	0.15	Objective	Rewards keeping prices attractive.
Waste weight	0.10	Objective	Penalises spoilage.
Volatility weight	0.10	Objective	Penalises price movement.
Annealing restarts / steps	24 / 1500	Solver	Quantum-inspired sampler budget.

02 · Parsed lines → 03 · Product model

Each product rolled up from the order text, then given its synthetic cost, stock and sensitivity assumptions.

Product	Lines	Units	Min ₹	Mean ₹	Max ₹	Baseline ₹	Cost ₹	Demand	Stock	Sensitivity	Spoilage
(Kaju) Cashew Nuts (W-210)	120	153	540	921	1355	883	620	15.2	56	-1.29	0.14
Avacado (Canadian)	48	70	145	162	175	160	115	6.9	28	-1.36	0.04
Chausa Mango (Uttar Pradesh)	12	18	200	223	240	220	155	1.8	5	-1.48	0.07
Dragon Fruit (Vietnam)	37	66	65	70	75	70	45	6.5	17	-1.54	0.14
Dry fruits Chocolate(Energy bar) (Salima)	42	87	180	203	220	200	140	8.6	26	-1.19	0.12
Dussheri Mango (Malyabadi)	3	3	150	200	300	150	105	1.0	5	-1.16	0.11
Imported Apples (Imported)	60	88	290	326	360	325	205	8.7	25	-1.19	0.06
Imported Apples (Queen) (Imported)	70	102	250	282	310	280	175	10.1	20	-1.05	0.03
Imported Blueberries (Canadian)	40	60	200	222	250	225	150	6.0	24	-1.76	0.07
Imported Orange (Mandarin)	44	65	180	202	220	200	125	6.4	15	-1.11	0.04
Imported Pears (Imported)	67	93	295	330	365	330	235	9.2	23	-1.51	0.07
Kashmiri Cherry (Kashmiri dark red)	14	14	440	492	540	490	325	1.4	5	-1.61	0.03
Langda Mango (Gujarati)	21	28	120	157	180	155	100	2.8	11	-1.38	0.10
Muscat green Grapes (Imported)	45	63	675	745	825	750	525	6.2	21	-1.25	0.13
Peru (Guava) (Indian)	47	95	80	96	105	95	60	9.4	41	-1.42	0.11
Rambutan Litchi (Muzaffarpur)	20	30	290	329	360	325	205	3.0	8	-1.13	0.05
Red & Dark Purple Plums (Kashmiri)	38	56	225	254	275	250	170	5.6	13	-1.09	0.10
Red king Pomegrante(Anaar) (Kashmiri)	39	86	120	134	150	135	95	8.5	27	-1.21	0.12
Shahi Litchi (Muzaffarpur)	2	2	300	300	300	300	210	1.0	5	-1.20	0.12

Cost, stock, sensitivity and spoilage are synthetic. Everything left of them comes straight from the uploaded orders.

04 · Price candidates & 06 · QUBO variable registry

One binary variable per product-and-price pair. Infeasible candidates are removed before the QUBO is built.

Variables x(i,k)

Feasible

Blocked by margin floor

Variable	Product	Price ₹	Index	Margin	Feasible	Demand	Profit ₹	Waste	Utility	Selected
x_1_1	(Kaju) Cashew Nuts (W-210)	795	0.901	22.0%	yes	17.4	3039	7.15	20.42	0
x_1_2	(Kaju) Cashew Nuts (W-210)	840	0.952	26.2%	yes	16.2	3558	7.21	30.66	0
x_1_3	(Kaju) Cashew Nuts (W-210)	885	1.003	29.9%	yes	15.1	4006	7.27	38.98	0
x_1_4	(Kaju) Cashew Nuts (W-210)	925	1.048	33.0%	yes	14.3	4355	7.31	44.22	x = 1
x_1_5	(Kaju) Cashew Nuts (W-210)	970	1.099	36.1%	yes	13.4	4700	7.35	48.87	0
x_2_1	Avacado (Canadian)	145	0.906	20.7%	yes	7.9	238	0.94	19.50	0
x_2_2	Avacado (Canadian)	150	0.938	23.3%	yes	7.6	265	0.95	26.52	0
x_2_3	Avacado (Canadian)	160	1.000	28.1%	yes	6.9	312	0.95	38.14	0
x_2_4	Avacado (Canadian)	170	1.063	32.4%	yes	6.4	352	0.96	45.88	x = 1
x_2_5	Avacado (Canadian)	175	1.094	34.3%	yes	6.1	369	0.96	48.92	0
x_3_1	Chausa Mango (Uttar Pradesh)	200	0.909	22.5%	yes	2.1	93	0.28	22.02	0
x_3_2	Chausa Mango (Uttar Pradesh)	210	0.955	26.2%	yes	1.9	105	0.29	32.33	0
x_3_3	Chausa Mango (Uttar Pradesh)	220	1.000	29.5%	yes	1.8	116	0.29	40.71	0
x_3_4	Chausa Mango (Uttar Pradesh)	230	1.045	32.6%	yes	1.7	125	0.29	46.56	x = 1
x_3_5	Chausa Mango (Uttar Pradesh)	240	1.091	35.4%	yes	1.6	133	0.29	51.05	0
x_4_1	Dragon Fruit (Vietnam)	65	0.929	30.8%	yes	7.3	147	2.03	21.12	0
x_4_2	Dragon Fruit (Vietnam)	70	1.000	35.7%	yes	6.5	164	2.06	40.86	0
x_4_3	Dragon Fruit (Vietnam)	75	1.071	40.0%	yes	5.9	177	2.10	53.35	x = 1
x_5_1	Dry fruits Chocolate(Energy bar) (Salima)	180	0.900	22.2%	yes	9.8	391	2.70	21.83	0
x_5_2	Dry fruits Chocolate(Energy bar) (Salima)	190	0.950	26.3%	yes	9.2	458	2.72	31.48	0
x_5_3	Dry fruits Chocolate(Energy bar) (Salima)	200	1.000	30.0%	yes	8.6	518	2.75	39.56	x = 1
x_5_4	Dry fruits Chocolate(Energy bar) (Salima)	210	1.050	33.3%	yes	8.1	570	2.77	45.30	0
x_5_5	Dry fruits Chocolate(Energy bar) (Salima)	220	1.100	36.4%	yes	7.7	616	2.79	49.93	0
x_6_1	Dussheri Mango (Malyabadi)	135	0.900	22.2%	yes	1.1	34	0.52	18.69	0
x_6_2	Dussheri Mango (Malyabadi)	145	0.967	27.6%	yes	1.0	42	0.53	31.42	0
x_6_3	Dussheri Mango (Malyabadi)	150	1.000	30.0%	yes	1.0	45	0.53	36.78	x = 1
x_6_4	Dussheri Mango (Malyabadi)	160	1.067	34.4%	yes	0.9	51	0.53	44.47	0
x_6_5	Dussheri Mango (Malyabadi)	165	1.100	36.4%	yes	0.9	54	0.53	47.58	0
x_7_1	Imported Apples (Imported)	295	0.908	30.5%	yes	9.8	882	1.34	21.66	0
x_7_2	Imported Apples (Imported)	310	0.954	33.9%	yes	9.2	970	1.36	30.65	0
x_7_3	Imported Apples (Imported)	325	1.000	36.9%	yes	8.7	1047	1.37	38.25	0
x_7_4	Imported Apples (Imported)	340	1.046	39.7%	yes	8.3	1116	1.38	43.75	x = 1
x_7_5	Imported Apples (Imported)	360	1.108	43.1%	yes	7.7	1197	1.39	50.10	0
x_8_1	Imported Apples (Queen) (Imported)	250	0.893	30.0%	yes	11.4	855	0.49	25.81	0
x_8_2	Imported Apples (Queen) (Imported)	265	0.946	34.0%	yes	10.7	965	0.50	35.34	0
x_8_3	Imported Apples (Queen) (Imported)	280	1.000	37.5%	yes	10.1	1062	0.50	42.91	x = 1
x_8_4	Imported Apples (Queen) (Imported)	295	1.054	40.7%	yes	9.6	1149	0.51	48.20	0
x_8_5	Imported Apples (Queen) (Imported)	310	1.107	43.5%	yes	9.1	1227	0.51	53.02	0
x_9_1	Imported Blueberries (Canadian)	205	0.911	26.8%	yes	7.0	386	1.48	19.44	0
x_9_2	Imported Blueberries (Canadian)	215	0.956	30.2%	yes	6.4	419	1.50	29.84	0
x_9_3	Imported Blueberries (Canadian)	225	1.000	33.3%	yes	6.0	446	1.51	37.74	x = 1
x_9_4	Imported Blueberries (Canadian)	235	1.044	36.2%	yes	5.5	469	1.52	42.73	0
x_9_5	Imported Blueberries (Canadian)	250	1.111	40.0%	yes	4.9	494	1.53	48.01	0
x_10_1	Imported Orange (Mandarin)	180	0.900	30.6%	yes	7.2	399	0.54	24.05	0
x_10_2	Imported Orange (Mandarin)	190	0.950	34.2%	yes	6.8	444	0.55	33.43	0
x_10_3	Imported Orange (Mandarin)	200	1.000	37.5%	yes	6.4	483	0.56	41.36	x = 1
x_10_4	Imported Orange (Mandarin)	210	1.050	40.5%	yes	6.1	519	0.56	47.07	0
x_10_5	Imported Orange (Mandarin)	220	1.100	43.2%	yes	5.8	551	0.57	51.73	0
x_11_1	Imported Pears (Imported)	295	0.894	20.3%	yes	10.9	655	1.37	23.95	0
x_11_2	Imported Pears (Imported)	315	0.955	25.4%	yes	9.9	791	1.39	35.75	0
x_11_3	Imported Pears (Imported)	330	1.000	28.8%	yes	9.2	876	1.41	42.27	x = 1
x_11_4	Imported Pears (Imported)	345	1.045	31.9%	yes	8.6	949	1.42	46.51	0
x_11_5	Imported Pears (Imported)	365	1.106	35.6%	yes	7.9	1030	1.44	50.83	0
x_12_1	Kashmiri Cherry (Kashmiri dark red)	440	0.898	26.1%	yes	1.7	190	0.15	20.94	0
x_12_2	Kashmiri Cherry (Kashmiri dark red)	465	0.949	30.1%	yes	1.5	211	0.16	31.96	0
x_12_3	Kashmiri Cherry (Kashmiri dark red)	490	1.000	33.7%	yes	1.4	229	0.16	40.18	x = 1
x_12_4	Kashmiri Cherry (Kashmiri dark red)	515	1.051	36.9%	yes	1.3	244	0.16	45.24	0
x_12_5	Kashmiri Cherry (Kashmiri dark red)	540	1.102	39.8%	yes	1.2	255	0.16	48.73	0
x_13_1	Langda Mango (Gujarati)	140	0.903	28.6%	yes	3.2	128	1.02	19.86	0
x_13_2	Langda Mango (Gujarati)	145	0.935	31.0%	yes	3.0	137	1.03	26.98	0
x_13_3	Langda Mango (Gujarati)	155	1.000	35.5%	yes	2.8	153	1.04	38.55	x = 1
x_13_4	Langda Mango (Gujarati)	165	1.065	39.4%	yes	2.5	166	1.05	45.92	0
x_13_5	Langda Mango (Gujarati)	170	1.097	41.2%	yes	2.4	171	1.05	48.72	0
x_14_1	Muscat green Grapes (Imported)	675	0.900	22.2%	yes	7.1	1070	2.49	21.08	0
x_14_2	Muscat green Grapes (Imported)	715	0.953	26.6%	yes	6.6	1260	2.52	31.52	0
x_14_3	Muscat green Grapes (Imported)	750	1.000	30.0%	yes	6.2	1406	2.53	39.07	x = 1
x_14_4	Muscat green Grapes (Imported)	790	1.053	33.5%	yes	5.9	1551	2.55	45.11	0
x_14_5	Muscat green Grapes (Imported)	825	1.100	36.4%	yes	5.5	1663	2.57	49.28	0
x_15_1	Peru (Guava) (Indian)	85	0.895	29.4%	yes	11.0	276	4.20	19.59	0
x_15_2	Peru (Guava) (Indian)	90	0.947	33.3%	yes	10.2	305	4.24	30.44	0
x_15_3	Peru (Guava) (Indian)	95	1.000	36.8%	yes	9.4	330	4.27	38.60	x = 1
x_15_4	Peru (Guava) (Indian)	100	1.053	40.0%	yes	8.8	350	4.29	43.84	0
x_15_5	Peru (Guava) (Indian)	105	1.105	42.9%	yes	8.2	368	4.32	47.93	0
x_16_1	Rambutan Litchi (Muzaffarpur)	295	0.908	30.5%	yes	3.3	299	0.33	22.13	0
x_16_2	Rambutan Litchi (Muzaffarpur)	310	0.954	33.9%	yes	3.1	330	0.33	30.96	0
x_16_3	Rambutan Litchi (Muzaffarpur)	325	1.000	36.9%	yes	3.0	357	0.34	38.51	0
x_16_4	Rambutan Litchi (Muzaffarpur)	340	1.046	39.7%	yes	2.8	382	0.34	44.04	x = 1
x_16_5	Rambutan Litchi (Muzaffarpur)	360	1.108	43.1%	yes	2.7	411	0.34	50.55	0
x_17_1	Red & Dark Purple Plums (Kashmiri)	225	0.900	24.4%	yes	6.2	343	1.05	23.96	0
x_17_2	Red & Dark Purple Plums (Kashmiri)	240	0.960	29.2%	yes	5.8	406	1.07	34.97	0
x_17_3	Red & Dark Purple Plums (Kashmiri)	250	1.000	32.0%	yes	5.6	444	1.08	41.21	0
x_17_4	Red & Dark Purple Plums (Kashmiri)	265	1.060	35.8%	yes	5.2	495	1.09	47.99	x = 1
x_17_5	Red & Dark Purple Plums (Kashmiri)	275	1.100	38.2%	yes	5.0	526	1.10	51.70	0
x_18_1	Red king Pomegrante(Anaar) (Kashmiri)	120	0.889	20.8%	yes	9.8	246	2.93	21.51	0
x_18_2	Red king Pomegrante(Anaar) (Kashmiri)	130	0.963	26.9%	yes	8.9	312	2.97	34.65	0
x_18_3	Red king Pomegrante(Anaar) (Kashmiri)	135	1.000	29.6%	yes	8.5	341	2.98	39.62	x = 1
x_18_4	Red king Pomegrante(Anaar) (Kashmiri)	140	1.037	32.1%	yes	8.2	367	3.00	43.18	0
x_18_5	Red king Pomegrante(Anaar) (Kashmiri)	150	1.111	36.7%	yes	7.5	413	3.03	49.45	0
x_19_1	Shahi Litchi (Muzaffarpur)	270	0.900	22.2%	yes	1.1	68	0.56	18.72	0
x_19_2	Shahi Litchi (Muzaffarpur)	285	0.950	26.3%	yes	1.1	80	0.57	28.62	0
x_19_3	Shahi Litchi (Muzaffarpur)	300	1.000	30.0%	yes	1.0	90	0.57	36.90	x = 1
x_19_4	Shahi Litchi (Muzaffarpur)	315	1.050	33.3%	yes	0.9	99	0.57	42.81	0
x_19_5	Shahi Litchi (Muzaffarpur)	330	1.100	36.4%	yes	0.9	107	0.57	47.57	0

05 · Optimization result

Latest saved run — exact solve next to the quantum-inspired solve.

Product	Baseline ₹	Exact ₹	QUBO ₹	Change	Margin	Demand	Sell-through	Profit ₹	Waste	Utility	Match
(Kaju) Cashew Nuts (W-210)	883	925	925	4.8%	33.0%	14.3	25%	4355	7.31	44.22	MATCH
Avacado (Canadian)	160	170	170	6.3%	32.4%	6.4	23%	352	0.96	45.88	MATCH
Chausa Mango (Uttar Pradesh)	220	230	230	4.5%	32.6%	1.7	33%	125	0.29	46.56	MATCH
Dragon Fruit (Vietnam)	70	75	75	7.1%	40.0%	5.9	35%	177	2.10	53.35	MATCH
Dry fruits Chocolate(Energy bar) (Salima)	200	200	200	0.0%	30.0%	8.6	33%	518	2.75	39.56	MATCH
Dussheri Mango (Malyabadi)	150	150	150	0.0%	30.0%	1.0	20%	45	0.53	36.78	MATCH
Imported Apples (Imported)	325	340	340	4.6%	39.7%	8.3	33%	1116	1.38	43.75	MATCH
Imported Apples (Queen) (Imported)	280	280	280	0.0%	37.5%	10.1	51%	1062	0.50	42.91	MATCH
Imported Blueberries (Canadian)	225	225	225	0.0%	33.3%	6.0	25%	446	1.51	37.74	MATCH
Imported Orange (Mandarin)	200	200	200	0.0%	37.5%	6.4	43%	483	0.56	41.36	MATCH
Imported Pears (Imported)	330	330	330	0.0%	28.8%	9.2	40%	876	1.41	42.27	MATCH
Kashmiri Cherry (Kashmiri dark red)	490	490	490	0.0%	33.7%	1.4	28%	229	0.16	40.18	MATCH
Langda Mango (Gujarati)	155	155	155	0.0%	35.5%	2.8	25%	153	1.04	38.55	MATCH
Muscat green Grapes (Imported)	750	750	750	0.0%	30.0%	6.2	30%	1406	2.53	39.07	MATCH
Peru (Guava) (Indian)	95	95	95	0.0%	36.8%	9.4	23%	330	4.27	38.60	MATCH
Rambutan Litchi (Muzaffarpur)	325	340	340	4.6%	39.7%	2.8	35%	382	0.34	44.04	MATCH
Red & Dark Purple Plums (Kashmiri)	250	265	265	6.0%	35.8%	5.2	40%	495	1.09	47.99	MATCH
Red king Pomegrante(Anaar) (Kashmiri)	135	135	135	0.0%	29.6%	8.5	32%	341	2.98	39.62	MATCH
Shahi Litchi (Muzaffarpur)	300	300	300	0.0%	30.0%	1.0	20%	90	0.57	36.90	MATCH

06 · QUBO model

How the business problem becomes an energy function.

Decision variables

93 binary candidate variables across 19 products.

One-hot penalty

A × Σᵢ (Σₖ x(i,k) − 1)² keeps exactly one price alive per product.

Business term

− Σ U(i,k)·x(i,k): lower energy means higher business utility.

Portfolio coupling

B × (average selected price index − 1.00)² holds the basket steady.

Minimum margin

Candidates under the floor are fixed to zero before sampling.

Quantum benchmark

Simulated annealing explores the energy landscape with one-hot-preserving moves.

Hardware caveat

QUBO-compatible for later D-Wave / QAOA / hybrid deployment; nothing here runs on a quantum processor.

07 · Benchmark

Quantum-inspired result measured against the exact optimum.

Metric	Exact classical	Quantum-inspired
Total utility	799.30	799.30
Average price index	1.0200	1.0200
Runtime	0 ms	0 ms
Identical price picks	19	19
Optimality gap	—	0.0000%
Expected profit	₹12,403 baseline	₹12,981

08 · SHAP global importance

Surrogate fit R² 0.9999 on the candidate utility function.

Price vs baseline7.065

Customer value2.161

Sell-through1.231

Forecast demand0.115

Expected waste0.100

Price sensitivity0.083

Price movement0.062

Spoilage rate0.059

Gross margin0.040

Stock on hand0.033

09 · SHAP local & 10 · LIME local

Pick a product to see why its price was chosen, and how sensitive that choice is.

Price raised by ₹43 to ₹925 — driven mostly by price vs baseline; margin 33.0%, sell-through 25%.

SHAP contribution to this decision

Price vs baseline5.121

Customer value1.206

Sell-through-0.837

Price movement-0.047

Expected waste0.030

Spoilage rate-0.028

Gross margin0.021

Price sensitivity0.021

Forecast demand0.019

Stock on hand0.017

LIME local sensitivity around the chosen price

Price vs baseline6.991

Gross margin0.069

Forecast demand0.135

Sell-through1.453

Expected waste-0.108

Customer value-1.847

Price movement-0.073

Stock on hand0.055

Price sensitivity0.034

Spoilage rate0.022

11 · 30-minute engine

What happens on every cycle.

01Refresh demand, stock and price inputs from the product model.
02Rebuild the candidate price ladder and drop anything under the margin floor.
03Solve exactly, then solve the QUBO with quantum-inspired annealing.
04Benchmark the two, record the gap and runtimes.
05Fit the surrogate and compute SHAP and LIME for every chosen price.
06Write the run, decisions, insights and knowledge chunks to the database.

Last run: 8/9/2026, 6:23:05 pm · trigger manual

12 · Dashboard summary

The workbook's headline numbers, live.

Products optimised

Expected profit

₹12,981

baseline ₹12,403

Portfolio price index

1.0200

band 0.98 – 1.02

Expected waste

32.3 units