 year and they break it into nine chunks and they say okay this is September 1st to October 15th and then they pick a characteristic week just arbitrarily pick the first week and then they optimize them if they tried to optimize everything the problem would be too large and it would become unsolvable in the finite amount of time so the decomposed it into a characteristic week and they optimize that one week and then they basically extrapolate that for the for the rest of that period so they generally solve the problem one week at a time and then basically copy and paste that information for a specific period ideally they should be able the goal is to solve the problem for every single day but we're just not there yet do you do crew scheduling as well not yet but it is in our horizon yes we have experience working in the crew area as well for a different area yes not not not ever like this yeah and just a reference point so most of these are so all of these are planning tools so these are used to create the schedule and then publish it at least six months out and as sky works is one that's more closer in as you as airlines sometimes have to make changes as you get closer to the date of operation yes yeah so yeah so our architecture is we actually have we have computational notes that actually do all the computations and then I don't want to say completely inclined but we have like a hybrid between dead and declined and a dedicated app server for the user interaction so the application is a web-based application but it's desktop and all the computations are done in remote what we call them computational notes which can be scaled up or down based off the needs of that airline okay I will hand it over to Sepone we'll walk you through the real case today thank you hope everyone enjoyed Sherman's speech now it's a boring time interesting about railroad is that it's much much much older than allies the industry itself was long long time ago you're telling the first time that people have something very close to railroad is like in 1500 when you actually have a horse pooling a carriage and actually you still have semi-track to actually let them track pooling your goods over the trail so it's still much better than using human or just pure horses a thing of change so much especially in 1800 when you actually have a steam engine right everyone have seen the steam engine like but yes does that actually revolutionize of exporting goods across the state across country with the real world let me show some more beautiful picture from our graphic design so before we go on to this you know trains people say train train trains all the time but what actually train train the composition of what is this anyone know what it's called locomotive locomotive locomotive engine local meaning from somewhere someone with no Spanish better than me but motive is that cause for motion right so you go from somewhere to another with some kind of engine so that really straightforward locomotive engines and basically the forecast of the top today but let me give some brief overview we will watch the video earlier you see how we come so far from 2000 actually our famous researchers happened to be also our speakers that do actually in the expert also he is an expert in network optimization and very first problem he worked with industry actually locomotive optimization why this is important simple is expensive it's very expensive so locomotive right now one of them would be created and our modern cost would be close to two to three million dollars for the locomotives but you say yeah this would be a very good engine and it never failed no it actually failed all the time they actually come to encounter through the reality so that's why watch how Monday for the airline makes sense you have a plan ally is a cuteing plan almost hundred percent of the time railroad yeah 50% you're lucky you know most of the thing actually in the railroad was actually unprecedented situation you don't know in the entire network for example the CSX is actually our main customer 20,000 miles or 35,000 kilometers half of the network was actually occupied by unscheduled trains the other half basically scheduled so when you have a mix of deterministic you know how it should go in with that in the domestic and mix together it actually gets created chaos to manage a chaos network you cannot rely solely on the plan and that's why we actually have our application to solve that problem what you see here can you guess what it is well you don't read on the top do you know what it is it's a yard so why we need yard yeah let's say it plays out where the action happened where the terminal people try to reassemble what is it the group of the cars actually come from google map so you zoom in and it's actually one of the biggest yard at CSX you see like thousand and thousand and car actually line up in several tracks and actually we also have an expert in the yard also simulation this is a type of the yard for hum yard there are two types of yard for for the railroad one is a hum yard one is a flat yard so what's that different hum yard usually the one is more efficient how it's more efficient because it's using gravity right you're actually pushing your cars over the hump and gradually slide out and integrate it without the cars and making something called blocks and then after that you're putting your locomotive in the front and pull it away you become trains and also uh so we're going to talk about why why this problem is important so us class one or basically the top top in the industry uh in us each one have between four thousand locomotive to eight thousand locomotives so you can see how much money they're spending they're not they're not actually afraid to spend uh they spend maybe two three years to replace about 10 percent of the locomotives so two billion dollars for them is nothing but because of their digging process right in the past they said hey this is also part of the capital they have to spend all the time but we want to share their mindset because if they can just save what charm and saving for the airline 10 percent let's say every year right we save the closer to what he did at least one or two million per day right and that is a lot a lot of money and this is interesting when we look at this it's actually in just a great cost 30 percent actually based on locomotives right it basically uh the the most important moving part of the railroad is locomotives and including our soul field but this is the last one is most interesting is that for all the locomotives they have only 50 percent have been used all the time and you will say why why they spend so much money and already half half of the time is anyone want to guess repair repair yes well that's one of the problem repair waiting waiting waiting waiting yes waiting for something to happen right yes and another interesting thing is that number of number of tracks number of tracks is limited your movement yeah i mean number of trips you need the locomotive okay so you said you don't have a lot of number of trips that's why yes yes so i think all this is is part of the problem but interestingly why they only 50 percent have been used because they don't know where their locomotives weren't warriors so why and i asked the same question why the thing is that look you have to remember that railroad actually have a very long history running for 100 and 130 years they're running through mainframe system a lot of locomotives actually won't equip with the gps so when they move around the yard and location you pretty much don't know sometimes you rely on people to call sometimes people just hiding and hiding locomotive is real why do you want to hide a locomotive this actually i think would have to guess but i still is fun to guess you know like basically each one getting their bonus based on on time departure you want your train you always depart on time right so what is it easier as a solution to do you're always holding your locomotives at your location you never be shocked at the locomotives that's why everything was said never can cause imbalance and you know demand supply just like what he did for the passenger demand supply same thing here along the network about 20 000 miles 35 000 kilometers sometimes it's more on the east east coast sometimes you're on the west coast so imbalance is happening all the time to make sure you counter the imbalance in general they're using repositioning right so basically you create your plan to move the locomotive against the ahead of the time of the demand of imbalance but if you don't know the prediction and everything so you just hold it you just keep it as much as you can so this is not critical at all we're just thinking through human minds okay so what we try to solve this is the product that we just started with cooperation with our partner we want to optimize the assignment and repositioning so assignment is nothing that nothing besides that you just put locomotives the right locomotive to the right trains and i'm going to explain more later and repositioning need that you move in from the location that actually have surplus to the one that actually need needed so shorter locomotives right and basically there's a lot of things happening in the real world all the time sometimes we delay and sometimes we broke down and everything we want actually the manager who managed the movement assignment and locomotives to actually just focus on the exception management what the exception management means i think doesn't go the way you think it would be right and basically we want to integrate this closely to operation so you see what we try to do is to help people in the field we want them to do the right things at the right time right what someone created is creating excellent plan that the people in the field can execute it this one is to help people in the field closer to the people who actually do this every day so if you're doing something suggesting something that people cannot execute you can create and have a walk right away and it's always changing second by second so improve on time departure and improve lonely utilization actually this is the biggest one so this is improve on time