Your SlideShare is downloading. ×
0
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Operation Research On Indian Railways
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×
Saving this for later? Get the SlideShare app to save on your phone or tablet. Read anywhere, anytime – even offline.
Text the download link to your phone
Standard text messaging rates apply

Operation Research On Indian Railways

2,965

Published on

Operation Research On Indian Railways

Operation Research On Indian Railways

Published in: Business
0 Comments
3 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total Views
2,965
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
359
Comments
0
Likes
3
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. OPERATION RESEARCH ON INDIAN RAILWAYS IIT-IBM Operation Research Workshop Bharat Salhotra Bharat.salhotra@sloan.mit.edu 8th April 2006
  • 2. INDIAN RAILWAYS 24*7*365 OPERATIONS
  • 3. Indian Railways Largest Railroad under single Management Revenue : US $ 12 Billion Surplus : US $ 2 Billion CAPEX : US $ 2 Billion Planning for ten-fold increase by 2012 Traffic Passengers : ~14 million /day Freight : ~2 million tons/day Manpower : 1.4 million 4/9/2006 Bharat Salhotra 3
  • 4. Indian Railways (Breakup of Revenue) k ul B 100 4% Misc. 50 al s er er en th O G 0 69% Freight 27% Passenger 4/9/2006 Bharat Salhotra 4 0% 10% 20% 30% 40% 50% 60% 70% 80%
  • 5. Infrastructure Track : 63,000 km. BG : 55,000 km. Traction Electric : 14,000 km. Diesel : 49,000 km. Signaling At stations : 5 types On sections : 6 types 4/9/2006 Bharat Salhotra 5
  • 6. Indian Railways: Complexity Complexity of Organization Divided into 16 Zones Investment Planning has been bottoms-up Complexity of Investments Investments merely shifts bottlenecks 24*7*365 Operations Change is the only constant Large # of Interdependencies Complexity of Operations Diversity of Traffic / Operations Mixed Operations 4/9/2006 Bharat Salhotra 6
  • 7. MICRO INTERDEPENDENCIES V ariation in booked speeds Variation in Slack Priority Groups +Impact of High + + Speed Turnouts Variation in Slope of train line + + Precedences ++ Variance in Commercial Speeds - V ariation in Throughtput Duration of Halts Variance in - HP/TL Ratios Variation in + Static time V ariance in achieved speeds + Speed + differential # of Passenger Train Halts 4/9/2006 Bharat Salhotra 7 Variance in Maximum Speed of Freight Trains
  • 8. Freight Trains on ST-BRC (22/3/05 to 31/3/05) Freight Trains on BRC-ST (22/3/05 to 31/3/05) 50 60 x= 287 40 x= 316 Frequency 40 30 σ=91 σ=327 20 20 10 0 0 120 180 240 300 360 420 480 540 600 660 720 780 120 180 240 300 360 420 480 540 600 660 720 780 M in Mi n Passenger Trains: BRC-ST Passenger Trains: ST-BRC 25 25 20 20 x=148 x= 137 Frequency Frequency 15 15 σ=44 σ=35 10 10 5 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 e 10 12 14 16 18 20 22 24 26 10 12 14 16 18 20 22 24 26 or Min Min. M All Trains: BRC-ST (22/3/05 to 31/3/05) All Trains: ST-BRC (22/3/05 to 31/3/05) 35 50 30 x=191 40 x= 178 Frequency 25 30 20 σ=173 σ=96 15 20 10 10 5 0 0 100 130 160 190 220 250 280 310 340 370 400 430 460 490 520 550 580 Mor e 0 0 0 0 0 0 0 0 0 10 16 22 28 34 40 46 52 58 M in. Min. 4/9/2006 Bharat Salhotra 8
  • 9. MACRO - INTERDEPENDENCIES Harmonization of Solution: link capacities Harmonization of Develop a model with three speed on a route objectives Reduction in Investment Harmonization Throughput of Train Speeds Model interdepen- 1 dencies + Reduction in Network Passenger Halts quality Take network - 2 effects into Reduction Introduction in marshalling consideration - of high- Limiting Passenger yards speed trains Market Assess bundle of 3 measures in a Shorter traveling Limiting Freight Reduced uniform way times + Market Revenues 4/9/2006 Bharat Salhotra 9 Source: McKinsey
  • 10. Need to understand Interdependencies for Network Planning High Opportunity Cost What is the cost effect What is the least cost & of changed passenger/ most Cost-Effective freight operating investment program? policies? Long Gestation What freight/ passenger flows can Capital Intensive be expected by 2012; Can new products be offered will the network on the network (e.g. Quick capacity be enough Transit Service) ; at what to cope? costs? 4/9/2006 Bharat Salhotra 10
  • 11. LONG RANGE DECISION SUPPORT SYSTEM SUITE OF TOOLS FOR STRATEGY PLANNING
  • 12. Long Range Decision Support System (LRDSS) Public-Private Initiative Conceptualized in 1995 Developed in 1998 Expanded in 2003 Powerful tool for pre-feasibility investment analysis for networks 4/9/2006 Bharat Salhotra 12
  • 13. LRDSS – Salient Features World-class in providing important desktop information for network planners and decision-makers for: investment planning financial impact analysis market analysis Uses Information as an Enterprise Resource for Decision Support 4/9/2006 Bharat Salhotra 13
  • 14. LRDSS : Salient Features Integrative Character: Interdisciplinary Network Oriented System wide Analysis Strong Decision Support “What-if” Analysis (“With/Without”) “Sensitivity” Analysis Information based & Data Driven. Iterative Evaluation Modular Design 4/9/2006 Bharat Salhotra 14
  • 15. LRDSS : Salient Features Customized GIS Interface Integration of different data by location Evaluate alternative routes Exhibit pattern of traffic flows Strategic tool Prioritize Investments by key years Position Services to optimize market share. Analyze Funds required by key year 4/9/2006 Bharat Salhotra 15
  • 16. Broad Structure of LRDSS Model Demand Analysis Project Traffic Market Identification Forecasting Analysis Supply Analysis Facility Traffic Performance Assignment Financial Cost Benefit Forecasting Analysis 4/9/2006 Bharat Salhotra 16
  • 17. LRDSS for Decision Support Annual Strategic Five Year PLANNING Plans PROCESS Level Plans Models to support different time Integrated horizons LRDSS ... BCAM DAF, FPM FPM 4/9/2006 Bharat Salhotra 17
  • 18. TRAFFIC ASSIGNMENT MODEL
  • 19. TAM-Conceptual design Replicate Shipper’s Behavior What commodity from where to where? Replicate Carriers Behavior What commodity, what route, what train type Non Linear Programming Solver MINOS 4/9/2006 Bharat Salhotra 19
  • 20. TRAFFIC ASSIGNMENT MODEL AS ORIGINALLY CONCEPTUALISED Production Amount O-D Demand Decomposition Shipper Sub-Module By Path from Consumption Algorithm User Optimized Production site Amount Path Construction Elastic Demand To Consumption Path Decomposition Aggregate Network Site Demand Function Traffic routed over Carrier Sub Module aggregate “network” as System-Optimized Shipper Sees it Fixed Demand Detailed Work Each shipper is non-cooperatively Minimizing landed cost Arc Flows, Arc Paths, Path Flows Model is a distribution, Modal split Path Costs Traffic Assignment Model 4/9/2006 Bharat Salhotra 20
  • 21. SYSTEM OPTIMIZATION Freight Network Equilibrium Model Model Shipper Behavior & Carrier Behavior Explicitly Accounts for the behavior the shipper & carrier (Frietz & Fernandez 1979) Non availability of road/inland waterways data prevented Shipper’s Assignment IR network controlled by single authority At equilibrium, Marginal cost of any path used is same! Transfer of flows to alternative path does not reduce cost 4/9/2006 Bharat Salhotra 21
  • 22. Carrier Model Assumption: Single carrier is in control Given (Oi-Dj)M faced by Indian Railways Path set Pkr(i..n1..n5…g1….t5…n3…n6..j) Arc-Path Incidence Matrix Congestion Cost over each arc: C=a+bXn Penalty Functions Shortage Variables Optimize Carrier Costs 4/9/2006 Bharat Salhotra 22
  • 23. Carrier Model Inputs A set of Origins & Destinations Oi, Dj A set of Commodities M1…….Mn Demand for Mn between Oi, Dj 4/9/2006 Bharat Salhotra 23
  • 24. Demand Modeling Different models used for different commodities GAMS Linear Programming Model Assigns traffic by minimizing transportation cost. “Furness” Trip Generation Model Generates OD flows based on movement pattern in the base year Factoring OD flows are projected based on growth rates. 4/9/2006 Bharat Salhotra 24
  • 25. Traffic Analysis Zones 4/9/2006 Bharat Salhotra 25
  • 26. Carrier Model Given A Set of Paths connecting Oi –Dj A Set of Links and Nodes comprising a path N1………Nx; n1……..ny A link /Node associated with a Cost Function CN1m=A+BXu Where A, B, U are constants, CN1m : Long Term Line Haul variable Cost A,B,u depend upon Type of Link Type of Train Type of Operating Policy U taken as 1 4/9/2006 Bharat Salhotra 26
  • 27. Cost Function Determination Congestion Curves obtained For each Train Type Each Link Type Each Operating Policy Set At different levels of Traffic Results converted into Cost congestion functions By Train Type, Link Type, Operating Rule set type 4/9/2006 Bharat Salhotra 27
  • 28. Traffic Assignment Cold Link with Micro Models for impedance Determination of Transit Times/ Cost per net tons For different trains At different levels of traffic Running on different links Speeds on Double Line Hilly Electrified MACLS Carrying different 80 commodities PRE618 70 PXE419 With different 60 operating policies PRE618 50 PPE412 PXE419 40 FHHEE5 30 Use of Simulation for PPE412 FHHEE5 20 FHHEE5 micro modeling 10 20 40 60 80 100 120 140 160 4/9/2006 Bharat Salhotra 28
  • 29. Baroda Surat Section Cost for different types of Passenger Trains 330 CONVERSION INTO COST Long Term Line Haul Variable Cost ( Rs.) per 310 FUNCTION 290 270 1000 GTKM Shatabdi 250 Slow Passenger 230 Rajdhani 210 Mail Express 190 170 1 35 44 53 61 80 4/9/2006 Bharat Salhotra 29 Trains per Day
  • 30. Traffic Assignment Operation Research based Freight Network Equilibrium Model. Objective function: Minimize Carrier Cost Assign OD flows on paths using least impedance. Σ congestion cost on links/nodes ) (= Each path consists of series of links and nodes. Path Cost = aggregated cost of traffic movement over each link and node. 4/9/2006 Bharat Salhotra 30
  • 31. Sub Modules of TAM Network Processor create a logical multi modal network access /egress links, transshipment, traction change points, nodes Output consists of Forward star data structure to be used as input to k path algorithm Path generator (K-Short) shortest paths between two O-D Pairs Input to Solver 4/9/2006 Bharat Salhotra 31
  • 32. Path Generator Path Generation Generate a set of 5 paths (max) from each Oi to each Dj Total TAZ’s 1456 Path generated only populated Oi-Dj Cells in Matrix Total paths generated = 40,000 K Short Algorithm 4/9/2006 Bharat Salhotra 32
  • 33. Network & Path Database Description Network databases Nodes are logical arcs Arcs are codified +,- for diesel *,~ for electric Path represented as follows 12 1 704.0 46 804.0 (Physical length= 704, links = 46) BL1257+ BD1093+ BN1252+ BD1966+ BC1238- BE1076*BN1237*BE1077~BN1239* BE1083~ BN1246* BE1218~ BN1405* BE1222~ BY1408~ BU1408~ 4/9/2006 Bharat Salhotra 33
  • 34. Sub Modules of TAM Carrier Input Processor Generates MINOS input file Represents full specifications of carrier model with unit costs specified as a ‘real valued’ function of path flows (non linear) Post processor Interprets MINOS solution file. Interfaces with GIS Query based GIS interface allows graphical display of bottleneck links, flows etc. 4/9/2006 Bharat Salhotra 34
  • 35. D a te : 3 0 th A p r il, 1 9 9 7 Low D a ta F lo w D ia g r a m R a ilL in e C ost M o d ific a tio n O p tio n s T r a ffic A s s ig n m e n t M o d u le S c e n a r io P r o je c t P in k B o o k R a ilL in e S e le c tio In v e s tm e n t F ile B a s e n D a ta N e tW o rk S a lv a g e D a ta R a ilN o d e M o d ifie r E le c tr ific a tio R a ilL in e F ile B a s e n P r o je c ts M o d ific a tio n d a ta R a ilw a y DAM & S ta n d a rd s MA M o d ifie d N e tw o rk M o d u le s C B A and N e w L in e / F ile D o u b lin g F F M o d u le M a in te n a n c P r o je c t O r ig in e C ost d a ta D e s tin a tio n L o g ic a l T r a ffic N e tw o rk S ig n a lin G e n e ra to r g P r o je c t D a ta E n try M o d if G e n e ra t D a ta O -D y e No In p u t G e n e ra to r R e p o rts L o g ic a l F ile s R o llin g N e tw o rk S to c k S o rte d C a lib r a te d P r o je c ts O r ig in d a ta G ro s s to D e s tin a tio Net Tons O -D s n T a b le P a th s C o n v e r s io n w ith o u t T r a ffic g e n e ra to r T a b le P a th s F a c ility No P r o je c t P a th D a ta A v a ila b le / P r o je c t Not R e p o rts P a th s C a lib r a tio n Yes F ile s A rc /P a th / S h o rta g e S e r ie s o f v a r ia b le s O u tp u t R a ilL in e P ro g ra m M IN O S R e p o rts to p re p a re F ile S p e c ific a tio n C a r r ie r M IN O S In p u t F ile In p u t F ile F ile D em and Rows R a ilN o d e A rc w is e R e p o rt F ile G ra d ie n t g e n e ra to r M IN O S In p u t C o m m o d it C a lc u la to r F ile y w is e ta b le P a th P e n a lty L in k C o s t V a r ia b le s C o s ts Lookup F ile A rc -P a th T a b le In c id e n c e M IN O S M in o s m a tr ix P ro g ra m O u tp u t F ile O r ig in C a p a c ity D e s tin a tio n R ow s F lo w F ile 4/9/2006 Bharat Salhotra 35
  • 36. Traffic Assignment Module Basic Inputs to the Model: I : Demand Side Existing and Future Traffic Commodity wise flows between pairs of points Traffic for 2006-’07, 2011-’12, 2016-’17 II: Supply Side Existing and Future Network Sections as well as their Characteristics Stations as well as their Characteristics 4/9/2006 Bharat Salhotra 36
  • 37. Methodology for Assignment Base Year: Assignment on Preferred Paths Future Years: Assignment on both Preferred & Shortest Paths Assignment with committed works Sequential/Simultaneous 4/9/2006 Bharat Salhotra 37
  • 38. Analysis of TAM Results Outputs Commodity wise traffic on each link. ODs that use a particular link. Lowest Cost Route path between pairs of points. Utilization of each Node Facility / Resource Planning at nodes These reports can be compared for alternative scenarios. 4/9/2006 Bharat Salhotra 38
  • 39. 4/9/2006 Bharat Salhotra 39
  • 40. GIS & LRDSS Avenue based Path Editor used to check paths generated by K-short transshipments, traction change, reversals create new paths via certain given stations User Interface to facilitate Data Analysis through Data browser Query Builder Chart generator 4/9/2006 Bharat Salhotra 40
  • 41. Path Editor to display/define routes 4/9/2006 Bharat Salhotra 41
  • 42. Path Editor 4/9/2006 Bharat Salhotra 42
  • 43. Bottlenecks in 2006-07 (Est. Capacity (188)) 4/9/2006 Bharat Salhotra 43
  • 44. 4/9/2006 Bharat Salhotra 44
  • 45. IMPACT OF LRDSS Focused Management on Long term planning Congestion Modeling Technology Process Improvements Market Analysis Pricing of products Quality of Service 4/9/2006 Bharat Salhotra 45
  • 46. PASSENGER OPERATIONS Challenge the Clouds!
  • 47. OVERVIEW Passenger Trains per day: 8000 Slow Passenger: 2500 Long Distance Mail Express: 1500 Contribute 90% of passenger revenues Local Trains:3000 MEMU Trains: 1000 4/9/2006 Bharat Salhotra 47
  • 48. OVERVIEW Long Distance Services/year: 303576 ~Trains per day = 1430 Investment = US $ 10 billion Revenues = US $ 2.4 billion Total Cars = 20,000 Rake Sets = 1351 Non Integrated Rake Links = 1000 Integrated rake Links = 351 Average Investment / rake set ~ US $ 6 mill. 4/9/2006 Bharat Salhotra 48
  • 49. Example: Non Integrated Rake Utilization of the rake: 31% Investment in Rake : US $ 4 million Mumbai Pune Departure: 5:40 AM Arrival: 9:15 AM Mumbai Pune Arrival: 10:00 PM Departure: 6:30 PM 4/9/2006 Bharat Salhotra 49
  • 50. Example: Integrated Link 4 Dharbanga 1e 1 3e Varanasi 5 2 5e 3 Pune Patna 6 Integrated Links Improve Utilization of Rolling Stock Utilization of rake: 85% 4/9/2006 Bharat Salhotra 50
  • 51. RAKE UTILIZATION Current Rake Utilization 52% Sensitivity: Improvement in rake utilization from 52% to 62% Saving in Rolling Stock Investment = US $ 1.3 b Additionally, revenue generation US $ 0.24billion Net implication = US $ 1.54 billion Improvement in Rake Utilization is critical 4/9/2006 Bharat Salhotra 51
  • 52. Current Approach Problem is impossible to solve “Trains will not run: services will be skipped for want of rakes” Hence, don’t even attempt it! Local optimization strategies Look at busy stations Analyze incoming trains Compare arrival timings and departures If close by, then match (subject to constraints) Length of trains Maintenance issues Train Configuration Result: Sub System Optimization but system sub- optimization 4/9/2006 Bharat Salhotra 52
  • 53. Proposed Approach ? OR + Project Management Service resembles a “project” Projects have a defined “start” and “end” Geography is an additional complication A train set is a resource Service utilize resources Objective: Complete all projects in time with minimum resources 4/9/2006 Bharat Salhotra 53
  • 54. Proposed Approach? Constraints Project start & end Location of starts & ends Location of resources Projects are attached to resources When Projects are completed, resource is available for next Project Resources are of different kinds 18 cars, 22 cars, 26 cars Projects seek specific resources 4/9/2006 Bharat Salhotra 54
  • 55. Proposed Approach? Use Traveling Salesman Analogy Consider 1351 salesmen are available Salesmen must depart & arrive at specific “stations” at “pre-specified time” Salesmen have pre-defined capacities Salesmen are not fully interchangeable Aim: Minimize no of salesmen 4/9/2006 Bharat Salhotra 55
  • 56. THANK YOU

×