Bridging the Gap in Humanitarian Operations Through Effective Partnerships Dr. Paulo Gonçalves Associate Professor – Università della Svizzera italiana (USI), Lugano Founder & Director – Master Humanitarian Logistics & Management – Master Humanitarian Operations & SC Management – Humanitarian Operations Research Center Research Affiliate – MIT Sloan School of Management Nairobi, September 17, 2014
Three Partnership Stories & Lessons Learned MASHLM & World Food Program (WFP) 2010 Optimizing Distribution of WfP’s Food Aid in Ethiopia MASHLM & United Children’s Fund (UNICEF) 2013/2014 Supply Chain Optimization of the distribution of mosquito nets in Ivory Coast in 2014 MASHOM & International Organization for Migration (IOM) 2013 MASHOM-IOM educational & capacity building partnership Different engagement & partnership models with different outcomes ! 1
Optimizing Distribution of World Food Program’s Food Aid in Ethiopia Bervery Chawaguta Paulo Gonçalves Logistics Officer Associate Professor WFP Ethiopia University of Lugano, Switzerland bervery.chawaguta@wfp.org paulo.goncalves@usi.ch WFP Ethiopia Addis Ababa November 3, 2010 2
Introduction Transportation: • Major cost component of humanitarian operations • Opportunity to increase cost-effectiveness • Opportunity to improve HO’s effectiveness Challenge: • Lack of systematic and reliable field data prevent organizations from optimizing distribution • Most existing optimization models applied to synthetic data 3
WFP Distribution on Somali Region WFP Ethiopia distributed 970,000 metric tones of food aid in 2009 • Transportation cost: US $65 million Primary transportation (from ports to hubs): • Cost: US$ 48 million Secondary transportation (from hubs to final destinations) • Cost : US$ 17 million Distribution context: • 3 Ports , 20 Hubs , 80 FDPs 4
Simplified WFP Distribution Example 370,000 90,000 Ports: • Djibouti P1 P2 (78%) • Berbera (13%) Hubs: • Dire Dawa (29%) • Jijiga (18%) • Nazareth (15%) FDPs: H3 H4 H5 • Nazareth (19%) 15,000 45,000 130,000 • Kombolcha (16%) • Jijiga (9%) • Dire Dawa (9%) • Mekele (7%) • Woreta (7%) F6 F7 F8 140,000 60,000 70,000 5
WFP Distribution Quantities Dire Jijiga Nazareth Kombol cha Mekele Woreta Supply Dawa P1 Djibuti 40,000 15,000 130,000 120,000 55,000 10,000 370,000 P2 Berbera 25,000 45,000 20,000 90,000 H3 Dire Dawa 10,000 15,000 15,000 5,000 15,000 60,000 Demand Supply H4 Jijiga 5,000 25,000 30,000 H5 Nazareth 10,000 5,000 5,000 20,000 40,000 F6 Kombolcha – – – – – – – F7 Mekele – – – – – – – F8 Woreta – – – – – – – Demand 75,000 75,000 170,000 140,000 60,000 70,000 – Note: Disguised quantities (supply and demand) to protect WFP’s confidentiality. 6
WFP Distribution: Rates, Supply & Demand Dire Kombol Jijiga Nazareth Mekele Woreta Supply Dawa cha P1 Djibuti 40 70 50 120 130 150 370,000 P2 Berbera 52 50 83 80 180 200 90,000 Cost (US$/MT) – H3 Dire Dawa 40 45 40 80 100 60,000 – H4 Jijiga 20 35 40 70 35 30,000 – H5 Nazareth 20 30 50 60 30 40,000 – – – – – – – F6 Kombolcha – – – – – – – F7 Mekele – – – – – – – F8 Woreta – Demand 75,000 75,000 170,000 140,000 60,000 70,000 Note: Disguised rates to protect WFP’s confidentiality. 7
Simplified WFP Distribution Cost Dire Dawa Jijiga Nazareth Kombol cha Mekele Woreta Supply P1 Djibuti 1600000 1050000 6500000 14400000 7150000 1500000 32,200,000 – – – P2 Berbera 1300000 2250000 1660000 5,210,000 – H3 Dire Dawa 400000 675000 600000 400000 1500000 3,575,000 – – – – H4 Jijiga 175000 875000 1,050,000 – – H5 Nazareth 200000 150000 250000 600000 1,200,000 F6 Kombolcha – – – – – – – F7 Mekele – – – – – – – F8 Woreta – – – – – – – Demand 3,100,000 3,850,000 9,010,000 15,250,000 7,550,000 4,475,000 43,235,000 Note: Disguised costs (from disguised rates and quantities) but useful as a reference to optimal values. 8
WFP Distribution: How much to ship? Dire Kombol Jijiga Nazareth Mekele Woreta Supply Dawa cha H3 H4 H5 F6 F7 F8 P1 Djibuti X13 X14 X15 X16 X17 X18 370,000 P2 Berbera X23 X24 X25 X26 X27 X28 90,000 – H3 Dire Dawa X34 X35 X36 X37 X38 60,000 – H4 Jijiga X43 X45 X46 X47 X48 30,000 – H5 Nazareth X53 X54 X56 X57 X58 40,000 – – – – – – – F6 Kombolcha – – – – – – – F7 Mekele – – – – – – – F8 Woreta – Demand 75,000 75,000 170,000 140,000 60,000 70,000 Note: Optimal quantities shipped are not know a priori, but can be solved using linear programing. 9
WFP Transshipment Formulation • Decision Variables: – X ij = Quantity shipped in arc ij, from node i to node j. • Objective: – Minimize total transportation costs • Subject to balance of flow constraints: – X 13 + X 14 + X 15 + X 16 + X 17 + X 18 = 370 P1 – X 23 + X 24 + X 25 + X 26 + X 27 + X 28 = 88 P2 – (X 13 + X 23 + X 43 + X 53 ) – (X 34 + X 35 + X 36 + X 37 + X 38 ) = 15 H3 – (X 14 + X 24 + X 34 + X 54 ) – (X 43 + X 45 + X 46 + X 47 + X 48 ) = 45 H4 10 – (X + X + X + X ) – (X + X + X + X ) = 130 – – ≥ 0, and ≤
WFP Distribution: How much to ship? Dire Kombol Jijiga Nazareth Mekele Woreta Supply Dawa cha H3 H4 H5 F6 F7 F8 P1 Djibuti X13 X14 X15 X16 X17 X18 370,000 P2 Berbera X23 X24 X25 X26 X27 X28 90,000 – H3 Dire Dawa X34 X35 X36 X37 X38 60,000 – H4 Jijiga X43 X45 X46 X47 X48 30,000 – H5 Nazareth X53 X54 X56 X57 X58 40,000 – – – – – – – F6 Kombolcha – – – – – – – F7 Mekele – – – – – – – F8 Woreta – Demand 75,000 75,000 170,000 140,000 60,000 70,000 Note: Optimal quantities shipped are not know a priori, but can be solved using linear programing. 11
Optimal Food Aid Distribution Potential Cost Savings Costs (Million US$) 65.0 65 60 10.3 -22% 55 4.1 50.5 50 0 2009 Actual Old Routes New Routes 12
Transport Cost Savings Average Rate (US$/MT) 240 GODE (3156) 220 200 KEBRIDEHAR (2116) 180 160 MOYALE (8) 140 BARE (9) 120 BOH (24) GELADIN (13) 100 80 HARSHIN (43) KEBRIBEYAH (306) DIRE DAWA (4384) 60 MEKELE (130) ADDIS ABABA (150) 40 AWASSA (159) 20 JIJIGA (1763) NAZARETH (2206) 0 Quantities (MT) 0 50,000 100,000 150,000 200,000 13
Transport Cost Increases Average Rate ($/MT) 240 220 DIHUN (1) 200 180 WARDER (2) 160 DEGEHABOR (689) 140 SEGEG (16) DANOT (2) 120 GERBO (21) 100 SHASHEMENE (55) DEBEWEIN (74) 80 SHILABO (68) 60 KOMBOLCHA (203) SHEKOSH (28) 40 MOJO (77) SHINILE (3208) 20 HOSSANA (84) AWBERE (34) GURSUM (3) 0 Quantities (MT) 0 50,000 100,000 150,000 200,000 14
Conclusions Transshipment optimization model can lead to significant cost savings: • Potential savings : – Existing routes: US $10.3 Mi 85,000 Mt – New routes: US $14.4 Mi 100,000 Mt • Clearly identified areas for improvement Significant commitment: • Shift from short- to long-term operations • Planning critical for success • Investment in new tools required • Systematic collection and analysis of data required 15
MASHLM-WFP Partnership – Failure Factors Lack of Senior Manager Support Senior managers interested in optimization tool and savings, but marginally involved in the process. Short-term Perspective Focus remained on short-term operations. No shift in focus or allocation of resources. Real, But Not Practical Application Focus on one year planning tool inadequate! WE tried to move into a shorter time horizon to influence current decisions, but no human resources were available. 16
S UPPLY C HAIN O PTIMIZATION OF THE DISTRIBUTION OF MOSQUITO NETS IN I VORY C OAST Irineu de Brito Junior Silvia Uneddu Paulo Gonçalves USP UNICEF USI (suneddu@unicef.org) (paulo.goncalves@usi.ch) (ibritojr@yahoo.com.br) 17
Malaria in Ivory Coast • Malaria is still endemic in CIV and is a priority of the National Health Development Plan 2012-2015 • Malaria is the leading cause of morbidity and mortality 43% of cases seen in health facilities 24% of hospital cases 26% of hospital deaths • From 2006 to 2008, utilization of LLINs has increased going from 3% to 14,8%, but only 26% of children under 5 received an appropriate malaria treatment. • To achieve and maintain universal coverage UNICEF planned a mass LLIN distribution campaign so that at least 80% of the population sleeps under the LLINs by 2015. 18
Modeling Tasks • Planning for the distribution of 12 million LLINs scheduled to take place in CIV in 2014. • Adopt quantitative project management tools to identify critical tasks and risk exposure (CPM, risk management). • Develop a linear programming model (transhipment) to identify constrains and possible bottlenecks • Review concept of operations to propose the most cost effective and efficient solution. 19
Project overview Ferkessedoug ou Bouak e Yamoussoukr o Abijdan San Pedro 20
SUPPLIERS AND PORTS LOCATION 21
Health District 22
Recommend
More recommend
