13
On the Algorithms of the Grid-Based EMMIL E- Marketplace Model Dr. Lívia Kacsukné Bruckner International Business School Inst. of Information Systems and Logistics Gábor Hermann MTA SZTAKI

On the Algorithms of the Grid- Based EMMIL E-Marketplace Model Dr. Lívia Kacsukné Bruckner International Business School Inst. of Information Systems and

Embed Size (px)

Citation preview

On the Algorithms of the Grid-Based EMMIL E-Marketplace

Model

Dr. Lívia Kacsukné Bruckner International Business School

Inst. of Information Systems and Logistics

Gábor Hermann

MTA SZTAKI

1. Combined negotiation with trading partners and logistics service providers

EMMIL E-commerce model

to integrate logistics

2. Choosing trading partners and 3PLs

Auction Algorithm for buyer-oriented marketplaces

1 .The buyer issues an RFP identifying the requirements 2. Sellers bid offering products. 3. The marketplace engine forwards the bids to the 3PLs who

place their bids for logistics services.4. The marketplace engine aggregates the offers from seller and

3PLs and forwards the best one to the buyer 5. The cycle is continued until the lowest accumulated cost is

achieved.

L

l Sj k

N

i

ki

lj

k

N

i

ki

lj

lj

kM

k

N

i

ki

ki

l lj

lj

QVZQFx

QP

1 11

1 1

/)(

1

min

Qik Purchased quantity of product i. from seller k.

Pik Unit-price of product i. at seller k. as a step

function of quantity Δk Discount given as a step function after the total

purchase cost at seller k.xj

l є {0,1} decision variable xj

l =1 offer j. of 3PL l. is selected as winner

The original Objective Function

Simplified Objective Function

to minimize:

j = 1

j = 1

i = 1

j = 1

i = 1

Pj,iQj,i + Vj Qj,i + Fj *

i = 1

Qj,i i = 1

Qj,i if

then i = 1

Qj,i

i = 1

Qj,i else

Great trick to obtain linear equations

j = 1

i = 1

t * Fj*yj,t +

t = 1

C

(Vj +Pj,i)Qj,t,i

Where C is the upper limit for

i = 1

Qi

and the value of the help variable yj,t must be 0 or 1

The constraints to solve the linear programming problem will be set so that Qj,t,i will be 0 if yj,t is 0

The bad consequence is

the big number of variables to solve the linear programing problem:

Qj,t,i instead of Qj,i and the additional yj,t

Haw to avoid combinatorical explosion?

• Idea:Diminish the number of sellers (Dimension j in Qj,t,i )

• Solution: Investigate only a limited number of “Best Sellers” (U), selected by the heuristic criteria “Every goods a delivered by the same seller”.Make the calculation parallel for S < U sellers in

US = U! /((U-S)!*S!) cases

GRID must be used to exploit parallel computing capacity

Implementation

Preparation, constructing the combinations

Call the LP_SolverCall the

LP_SolverCall the LP_Solver (i)

Collect and compare the results

1

2

US

Pilot Implementation using the P-GRADE Portal

Solution details: http://www.sztaki.hu/~ghermann/Szemelyes/EMMIL_2007/

Project 1.:Test the existing EMMIL model on the P-Grade portal for the following scenarios: Enhance the basic version with capacity (upper) limits for products at suppliers. This involves the following:Add an array to the data structure that contains the capacity limitsInteger Limits[k,i] (k=1..M, i=1,..N) Change the generator programFill up the Limits table putting a normal distribution noise to the required quantitiesChange the prefilter program to cope with quantity limitsPass the relevant limits to the PS jobsPublish all data that has been generated, in a nice customized format. Change the PS job programFacilitate receiving the limits from the input fileGenerate additional conditions for the solvers to take quantity limits into consideration. Call the solver only if the selected sellers together have enough of all products otherwise generate the relevant error code.Enhance the collector program to handle errors properly.Test your solution. Compare the execution times with different granularities.Document your work.

Project 2.:Test the existing EMMIL model on the P-Grade portal for the following scenarios: Enhance the basic version with an extra feature to allow discounts on combination delivery from different suppliers. This involves the following:Change the user interface to allow entering combinations of seller indexes and discount percentages. Add an array to the data structure to store it.Integer Combination discounts[s1, s2,,..percentage] Change the generator programChange the prefilter program and make sure that the combination where discount is specified will be passed to the PS jobs.Pass the combination discounts table to the PS jobsPublish all data that has been generated, in a nice customized format. Change the PS job programFacilitate receiving the discounts from the input fileCheck if there is a discount for the current seller combination and change the fix cost of the transportation when generate the input for the solver.Test your solution. Compare the execution times with different granularities.

Project 3.:Test the existing EMMIL model on the P-Grade portal for the following scenarios: Modify the basic version and create a benchmark application where optimization is done only for prices, transportation costs are added after the best solution is found. This involves the following:Change the generator programChange the prefilter program to find candidates for best prices onlyPublish all data that has been generated, in a nice customized format. Change the PS jobsModify the data that is passed to the solver excluding data about transportationChange the collector programAfter finding the best allocation calculate the total cot by adding transportation costs to the purchase costTest your solution in parallel with the basic solution using the same data and publish the comparison of the results.