4 th Internationa conference on Information Systems, Logistics and Suppy Chain Quebec City August 26-29, 2012 Modeing a Scenario-based Distribution Network Design Probem in a Physica Internet-enabed open Logistics Web Heia Sohrabi Lava University, CIRRELT, Canada Waid Kibi BEM- Bordeaux Management Schoo, CIRRELT, France Benoit Montreui Lava University, CIRRELT, Canada Research Chair in Enterprise Engineering, Canada August 29 th, 2012
Presentation pan Current distribution context Physica Internet initiative: Introduction Research methodoogy Future business environment modeing Distribution network design generation Mathematica mode o Generic mode o Context-driven mode Fina remarks and future work 2
Current distribution context Cosed distribution system (Private Distribution Network; PDN) Cost-service trade offs Addressing changes in the business environment (e.g. customer expectations, economic instabiity and energy crisis) Coaborative distribution system (Shared Distribution web; SDW) 3
Physica Internet Initiative: An Introduction What is Physica Internet? An open goba ogistics system founded on physica and digita interconnectivity through encapsuation, protocos and interfaces What is Physica Internet goa? To reduce by an order of magnitude the environmenta, economic and socia efficiency and sustainabiity of the way physica objects are moved, stored, reaized, suppied and used http://physicainternetinitiative.org 4
The Physica Internet aims to enabe an efficient and sustainabe Logistics Web Research question: Logistics Web How to design distribution networks expoiting the open distribution web? Set of openy interconnected physica, digita, human, organizationa and socia agents and networks aiming to serve efficienty and sustainaby Research goa: the ogistics needs of peope, organizations, territories and society To investigate Reaization the Web performance improvement (both from economic Suppy and Web environmenta Reaizing products perspectives) achievabe by switching from current Suppying distribution goods context Interconnected to Physica open production, Internet enabed context. personaising & retrofit centers Distribution Web Depoying, storing products Interconnected open warehouses & distribution centers Mobiity Web Moving goods & peope Interconnected open unimoda & muimoda infrastructures, vehices, hubs and transits Interconnected open suppiers and subcontractors Service Web Enabing and sharing access and usage of services rendered by goods & peope Interconnected open users and service providers 5
Kibi, W. and Marte, A (2012) For company 1 Research methodoogy company m Future business scenario generation (Monte Caro approach) Distribution Network ODW Distribution Network design SDW generation Distribution Network design generation PDN at each distribution context design generation each distribution context at each distribution context Distribution Network Distribution Network performance investigation Distribution Network performance investigation (simuation) performance investigation (simuation) (simuation) ODW SDW A singe fina design in PDN context 6
Future business environment modeing Appying a nove scenario panning approach based on evoutionary paths (Kibi and Marte, 2011). An evoutionary path is a possibe view of the future characterized by the future-shaping variabes. Future-shaping variabes defined: Demand-market, Avaiabe capacity, (incuding production, storage capacity) Network cost, (incuding investment cost, operationa cost) 7
Future business environment modeing Three defined evoutionary paths I. Sustainabiity-unconscious II. A pessimistic view of the future word Based on the unstabe situation of severa economies in the word and the scrambe energy scenario described by She (2009). Sustainabiity-prone A word that is conscious about the need to become more sustainabe, yet is sow in moving towards sustainabiity. anticipated to emerge in the parts of the word, mosty in North American and Western European deveoped countries. Moderate improvement in economy and oca actions addressing energy security, in ine with the bueprint scenario eaborated by She (2009) III. Sustainabiity-engaged. A more optimistic view Inspired by the visions of the Physica Internet (PI) (Montreui, 2011). 8
Future business environment modeing future shaping variabes evoution through each evoutionary path Sustainabiity-unconscious Sustainabiity-prone Sustainabiity-engaged Using a Monte Caro procedure For each evoutionary path For each future-shaping variabe Cost of investment Panning horizon A set of scenarios wi be generated Accessibe capabiities Panning horizon This procedure wi be repeated for each company case with different characteristics independenty from each other. Operationa cost Panning horizon 9 Demand Panning horizon
Distribution network design generation Strategic reengineering process A reengineering cyce Yeary/quartery panning period t T Weeky/Daiy working period τ t(τ) Panning horizon Panning horizon 10 Transportation cost in a scenario Demand cost in a scenario Avaiabe capacity cost in a scenario Panning horizon
Schematic representation of three distribution contexts PDN Panning horizon SDW Panning horizon ODW 11 Panning horizon
Faciities fixed and usage costs in different contexts Initia state Do not use the site Use the site Decision Fixed cost Decision Fixed cost Usage cost Current faciity Own (o) o c L L Cose Transfer and cosing cost a Use during a the horizon T Capita recovery a Operating cost a L c Rent/Use service (r) r c L L Stop for a period Cosing cost pus ease penay Renew renta contract for a period t T a, t T tt Lease cost Operating cost a, t T a, t T Potentia site L p Purchase/ Buid o p L L Do not use Zero Rent Do not use Zero r p L L Open and use during a the horizon T Agree a renta contract for a period t T 12 Setup cost pus capita recovery a Setup cost pus ease cost Operating cost a Operating cost a, t T a, t T
The generic mathematica mode E o L NOP Max R x, a x a x a x a x t tt r o L L a x a x (1) Subject to: Customer-DC assignment constraints: x st x L, s S and t T (2) DC state constraints (for rented faciities): x x 1 r L, t T (3) 13
DC opening/cosing constraints: The generic mathematica mode continued x x x x ( t1) 0 r L, t T (4) DC state constraints : x t T x x 1 DC state constraints : x x x 0 0 o L (5) o L (6) Non negativity constraints: x, x, x, x, x, x {0,1} st L, s S, t T (7) 14
The generic mathematica mode continued Where for a given panning period t T x t T,, R, Max [( u ( ) F ( ) t p ss ps ps p, v, p, ss ' p,, ( LS ) p, g k f pv p e e ( u ( ) f ( )) F ( )) ps ps ps F f I Z ' p p pv F ' )] p Tota income externa sourcing cost Transportation cost Inventory cost Recourse capacity acquirement cost (8) 15
The generic mathematica mode continued Subject to: Capacity constraints: I ( ) c ( F ( )) b ( ) Z ( ) p p p p ss ps t L, T, (9) Customer assignment constraints: F ( ) d ( ) x t t p, T ps p, T ps st L, s S, (10) Inventory/ product fow constraints: I ( ) I ( ) F ( ) F ( ) p p ( 1) ( LV ) p ( LS ) p t p P, L, T, (11) 16
The generic mathematica mode continued Initia inventory constraints: I p ( ) 0 0 Demand satisfaction fow constraints: e d ( ) F ( ) F ( ) d ( ) ps ps L ps ps Recourse capacity constraints: Z ( ) b ( ) x p P L T 1,,, (12) t p P, s S, T, (13) t L, T, (14) Production capacity constraints: F L pv ( ) b ( ) pv Integraity and non negativity constraints: e F ( ), F ( ), F ( ), Z ( ), I ( ) 0 pv p ps p 17 t p P, vv, T, (15) t p P, vv, s S, ( L S), T, (16)
Private distribution network Context-driven design modes Current context(cosed) The design mode in PDN is simiar to the generic mode (1)-(16) The optima design soution is denoted by and the vaue of this soution * is denoted by E NOP x. x * 18
Shared distribution web Context-driven design modes Current context(coaborative) Group of M partnering companies Each company designs a distribution web maximizing its own performance subject to joint capacity restrictions for each shared faciity The sets of ocations dedicated to each company The set of shared faciities L m m M m L m m ' \ ' Vm, Lm, Sm The portfoio of avaiabe resources * The optima soution and its vaue L L m m x * m E NOP x m 19
Context-driven design modes Open distribution web context Companies mosty depoy their products independenty from each other even though coaborative behavior such as in SDW is possibe. Dynamic avaiabe capacity b ( ) Two additiona set of resources : A private faciities operated by other companies L m Pubic faciities whose storage capacity is competey open to the market, denoted by Uimatey, the open distribution web incudes both companies, and third-party ogistics providers openy offering distribution storage capacity. w L m L L L L m m m m w 20
Fina remarks and future work Fina remarks: Sampe Average Approximation (SAA) (period samping, scenario samping and/or aggregations) Exact approach (OPL-CPLEX) / Tabu based metaheuristics Future work: To vaidate the mode on a representative case Comparison between the three aernative designs in different distribution contexts studied using a arge set of scenarios and a various set of criteria such as: deivery speed, eco-efficiency, resiiency, etc. 21
Your questions and comments are wecome. 22