Department of Computer Science and Automation Big Data Public Lecture Series Data Based Decision Making in Manufacturing Supply Chains N. Viswanadham INSA Senior Scientist Computer Science and Automation Indian Institute of Science, Bangalore July 3, 2014 Contents Manufacturing Supply Chains Example: Changing Face of Auto & Logistics Industries The Big Data Ecosystem The Procurement Process: Data Based Decision Making Conclusions 1
Manufacturing Supply Chains Input-Output model of a Manufacturing System 2
Integrated Manufacturing Supply Chain Network B2B Logistics Chain B2C Logistics Chain i i i i i Supplier OEM Distributor i i i Service Center i i Customer i Logistics Integrated supply chains are a network of Suppliers, Contract manufacturers, Distributors, Retailers, Logistics providers, Repair & Maintenance providers. Multi Tier Supply Chain Network An auto supply chain has about 15000 companies in its SCN 3
Recent Advances:Internet of Things,.. IoT technologies can be categorized into Tagging things, Sensing things and Embedded things. The tagging things provide item identification, things can be connected to the databases. The sensing things enable us to measure and detect changes in the physical status of our environment. The embedded things yield information about the status of the embedding object. Cyber Physical Systems Systems of Systems Network of Networks Analytics 1.0: Decision Making using Internal Data Several long term and short term decisions are made Sourcing: which country & from whom Demand estimation using sales data How much to manufacture, inventory levels at various places to match the demand ERP, APS, TMS,WMS etc make decisions analysing internal data: sales, shipments, inventory, etc. Control using PLCs, Robots, BPOs etc. Monitoring equipment for preventive maintenance using IOT 4
Integrated Information Systems Customer Orders Purchase Orders SUPPLIERS ERP Finance HR MRP Sales History Manufacturin g Schedule Production Picks Duty Completed Inter-site Transfers POD Pick Detail Receipt Detail Customer Orders Global Logistics Customer Orders WMS POD Vehicle Routes Orders for Routing Exceptions Carrier Discrepancy TMS POD POD Customers ASNs Demand Forecasting APS Manufacturing Scheduling Demand Planning Production Planning Inventory Summary Inter-Site Transfers YMS Load & Dock Detail EDI Biddin g Carriers ASNs ASNs ASNs Sudden and Synchronized Trade Collapse Trade flows dropped by more than 20% from 2008Q2-09Q2. The synchronization was due to the connectivity of global supply chains that reacted just in time to the collapse in demand 5
Big data Enabled Business Processes Big Data in Supply Chains Procurement: Supplier & Logistics provider selection, Inventory management Dispersed Cognitive Manufacturing: Embedded Machines, Smart parts, Cognitive PLCs Distribution & Retail: Warehousing, B2C Logistics, Recommender systems Service Chains: Logistics networks, Repair & Maintenance of Machines, Trucks, etc., Traceability and Product recalls Risk Mitigation Examples: Changing Face of Auto and Logistics Industries 6
The Auto Industry is going through Resource Revolution Cars are the second biggest capital expenditure we make. They are parked 96 % of the time. Average occupancy is 1.6/5. In the rest 4% is spent looking for parking, waiting at the traffic lights and in driving Machine that Changed the world Good drivers are in a minority. Millions of accidents and deaths a year & 33 % of drivers didn t touch the brakes before collision. Anti-collision technologies face liability issues if control of a car is taken away from the driver. Mobile Application Start-ups Making History BMW and Daimler say they are transportation companies Zipcar acquired by Avis in 2013 for $500 million, lets people rent cars by the hour in major cities; each Zipcar replaces 21 cars. Uber lets people summon a car and driver via a smart phone app. Google invested $258 million in Uber. mgaadi is bringing Uber-like convenience to auto rickshaw riders UberX in Bangalore, Delhi and Hyderabad, directly competes with startups such as Ola Cabs and TaxiForSure. RelayRides and Getaround provide marketplaces where one can rent their cars to others. Inrix gathers location information from millions of mobile devices and feeds information on traffic flow and optimal routes to drivers. Real Disintermediation happening 7
Uber for Logistics is Happening in Asia Gogovan and Easyvan provide a peer-to-peer app that connects van drivers with individuals or businesses who need their stuff shipped quickly Trucking firms use data from new sensors monitoring fuel levels, location and capacity, driver behaviour, etc. in their optimization. The goal is to improve the company s route network, lower fuel cost, and decrease the of accidents Uber Isn t a Car Service. It s the Future of Logistics Genpact: Control Tower for Penske Genpact Orchestrates the logistical services of Penske Genpact BPO workers in India and Mexico Check the customer s credit status and arrange necessary permits. Assign Trucks and Drivers based on driver s choice and also the maintenance record of the truck If the truck gets stuck at a weigh station for permits, the BPO staff transmit the necessary documentation. After the trip, the driver s log is shipped to a Genpact facility Penske processes lot of info: numerical, text, voice, past records of 15K trucks and equal number of drivers A mobileapp in India is in need 8
Google has Created a Race in Driverless Cars. Google s driverless car has a license to operate in California, Florida, and Nevada. Driven 700,000 miles & No accident. They use 360 deg sensors, lasers, learning algorithms and GPS AI software in the Google car learns from every experience of every car and will generate a real-time map of road conditions Cars could be managed as a network in the future Nissan & Daimler are committed for driverless cars by 2020. Rio Tinto s driverless trucks moved 100 million tonnes Huge implications in social, industrial & military sectors Service, Maintenance, and Repair 75% of power plants run on natural gas, oil, coal or nuclear Using Big data analytics, Aircraft will tell maintenance crews the status and which parts need replacement GE can predict failure of gas turbines weeks in advance (IOT) A shift from current practice of Maintenance being carried out on a set timetable or reactively GE invested $105 million in Pivotal, a Big Data company formed by EMC and VMware. 9
Retailing: Disruptive Changes Big Data in Supply Chains Retailers watch the shoppers in the store, where they go, in what order and understand how all these map to actual sales. Recommender Systems suggest to consumers products based on their browsing, searches and earlier purchases Netflix uses recommender system for each subscriber. Target predicted pregnancy in a Teen based on her buying patterns. Focus shift from Sales & Marketing to Predictive Analytics using Industry Knowledge, Consumer preferences, Connections with the Stakeholders, Social media analytics Privacy is at stake??? Data science is the next big thing in Agriculture Monsanto acquired Climate Corporation, maker of a software platform that crunches weather-related data to help farmers grow crops more effectively for $930 million. 10
Implications for SCNs Demand for Services not just products: Power by Hour New services such as information networks & protocols for roads and control of traffic are needed Happening in mines and military Big Data Ecosystem 11
Big data should aid in Decision Making that results in Desired Business Outcomes "What is the most desired business outcome and what marriage of data and algorithms gets us there?" The Big Question what data from suppliers, customers, governments, and local & economic environment should one collect and analyze What data you analyze every day, every week, every month. A framework is needed to answer this question Examine a close alignment between NYSE and London Stock Exchange indices and the amount of solar energy hitting the earth. One might draw a conclusion that the amount of solar energy drives stock prices based on this data. It just happened to be a coincidence over a relatively narrow window. Make Sure It s Relevant. India Logistics Project N. Viswanadham 12
The Basic Ecosystem Institutions Resources Delivery Services Big data Service Chain Regulatory Bodies Legal Licenses & Privacy Issues Central /State Governments Citizen Groups, Social Activists Business Organizations AI Based Decision support systems BPO Decision Making tools Communication Tools Feedback & Correction Training Staff Delivery Technologies & Mechanisms Institutions Big Data Ecosystem Data Service Chain Vertical based Resources Human Resources with new skill sets Cloud & other Storage Resources Software Clusters Social Media R&D and Educational Institutions Wireless and Smart phone service providers Content Identification Acquisition Analysis Control Action Data Marshalling 13
The Procurement Process: Data Based Decision Making B2B Procurement Strong ties with Trusted suppliers Total landed cost Focus on supplier ecosystem not just product price & quality 14
The Basic Ecosystem Institutions Resources Delivery Services Infrastructure Supply Chain Investment Climate Customs, Export & Trade Other Govt. Regulators Quality Control & Environmental Issues Social, Legal and Privacy issues, Labor Unions Logistics & IT companies Delivery Channels Institutions Infrastructure: Ports, Airports, Roads, Industry Clusters Decision Making Tools, BPO, Control Towers IOT and Supply Hubs Delivery Service Mechanisms Procurement Ecosystem Resources Cloud Social Media, Recommender systems Human, Financial & Natural Resources Location Factors Procurement Chain Manufacturing Distribution Logistics Suppliers 15
Supplier Selection using Transaction Costs Delivery Shipping, Inventory, Asset specific Hard & Soft Infrastructure Resource Asset Specific Clusters, Human, Financial, Power Institutions Taxes, Tariffs, SEZs, FTAs, Social groups Transaction Cost Supply Chain Production, Quality, Transport Coordination Costs Broker fees Conclusions Our framework identifies the data to collect and analyze to make the needed decisions Data formats need to be standardized for easy collection Attention is needed in creating apps for disintermediation The Indian truck market, where 80 % of operators own less than 10 trucks & Majority of them are owner-drivers with a single truck and is organized by transport middlemen or goods booking agents Same is true for SMEs Farmers,Commission Agents,Traders, Industries, Retailers, Wholesalers, Consumers, Mandi Staff form the social network. Applicable to service value networks and public networks such as infrastructure, public health and food security. Talent is working on other s problems. Attention to Indian problems is the need of the hour 16