On the topic of simulations, I've been learning about Modelica, which is a engineering simulator. Basically, many real world devices can be described using system dynamics and their behaviour codified as differential equations. Modelica allows us to build such models, including mixed-domain models (electrical, mechanics, thermal, chemical, etc.), then the Modelica compiler produces highly optimized C++ code you can run to perform the simulation. It's heavily used in the car industry in Europe, but not widely known in academia.
Context: I'm researching a new book on ordinary differential equations through various real-world simulations, so like an ODEs theory book, but actually applying the concepts to real world systems. Get in touch by email if you're interested in seeing advanced preview of some of the projects that will be developed this summer. Contact info in profile.
Ivan - Since I regularly recommend your maths and physics books (and use and do research with Modelica)- very very interested in this. Good luck and please let us know when it can be pre-ordered.
Ah no, I'm very far from the "book draft" phase — I'm more at the "Hello world" stage ;) I've been scoping out the hands-on projects first (e.g. simulate electric motor, heater, AC unit, heatpump, etc). I think adding the ODE theory will be the easy part, so I'm focussing first on the real-world projects/applications, which might be interesting to release as standalone blog posts to make the company blog look alive https://minireference.com/blog/ ;)
Assuming you mean Simulink rather than Matlab its not quite a direct alternative. You can use Modelica and Simulink for modeling the same type of problems but they use different modeling approaches. Simulink uses cuasal modeling, where signals flow via connections between individual blocks, transmitting values of individual variables from the output of one block to inputs of another block. Modelica uses acuasal modeling - in Modelica, individual components of the model describe the equations directly and by interconnecting individual components, the systems of equations become connected.
If you are looking for a direct open source causal modeling alternative to Simulink, then SciCos (www.scicos.org) is worth a look. I would recommed playing with both Open Modelica and SciCos to see which is the best fit for you.
>Chevron paid Maxis $75,000 for a prototype of a refinery simulator
> I came up and said, “How about, these concepts, principles, and their relation?” and they say, “Oh, John, you screwed up, look at – this is wrong.” And I say, “Okay. Help me understand what to do about that.” They did, and they taught me. And all I was doing was field work. I was asking questions and listening carefully. And I’d say, “What about this?” I’d look at a photograph of their oil refinery, and I’d say, “What’s this big pile of sulfur doing here?” “Oh, well you weren’t paying attention to the chemistry, John.”
That $75K really caught my eye. I know that is ~$140K in today's dollars, but it still seems like a lowball number given: a) the amount of domain knowledge and research Maxis developers needed to do, b) that this game was intended to be a training tool for not just actual plant operators but non-technical staff, and c) Chervon's market cap and (I assume) massive budget for doing whatever they want.
Of course Maxis and its biz-sim division were relatively new, and the value of games and digital simulation much different than today, but I can't imagine that kind of corporate project today being agreed to for anything less than $300K (in 2020 dollars)
You're right, a similar project these days would probably going into the millions. You'd have all kinds of project managers and 'scrum leaders' and QA and who knows what else. And people probably billed out at $60/hr back then rather than $300.
Smaller teams, simpler products, it was a different time.
The article hints that the project started around the time Maxis were closing on a $10m investment from a VC interested in the market potential business simulation and Delta Logic were to be acquired as part of the deal. So there must have been a lot of motivation on the side of the developers to do the deal at any cost. Chevron apparently publicly described it as a prototype, and I suspect the project documents did as well, along with ambitious outline proposals for how it could be expanded at much greater expense if the pilot was successful.
Cities Skylines would be the best bet at a modern equivalent. They do still make SimCity games, but they're pretty far removed from the mechanics of the old game.
Oxygen Not Included might be close to what you are looking for. It's based in space, but its basically a systems-based game for managing a space colony. Includes managing your air (how much oxygen, handling co2), waste, water, spread of disease, and more.
Man I love that game. It has a very clever automation/sensor system as well. You can rig doors to open if they sense water, heat, gasses, pressure, and a bunch of other criteria. Wiring in NOT/AND/OR gates. Timers.
The thing with factorio is the green power generators still have pollution associated with it, so I always just elect to use the more powerful coal generators and beef up my defenses.
To be clear: Solar panels in Factorio produce no pollution, but metal smelting produces huge amounts of it, and you have to smelt to do basically anything in the game.
Direct reduction of iron ore with Hydrogen can eliminate most of the process emissions of smelting: https://en.wikipedia.org/wiki/Direct_reduced_iron . Production of secondary steel (recycled steel) using electric arc furnaces also barely has process emissions.
I've never heard of this before, so correct me if I'm wrong, but that wikipedia article makes it look like 2Fe2O3 + 6H2 -> 2Fe + 6H2O. Commercially hydrogen is produced via CH4 + 2H2O -> CO2 + 4H2.
Combined you're just converting methane into CO2 and H2O as if you burnt it? Is it actually the case that this ends up needing less hydrocarbons than other routes for some reason?
I think[1] typical plants steam reform natural gas to CO and H2 and use that to reduce iron ore below the melting temperature of iron. Which makes it energy efficient.
In the US, no new blast furnaces have been built in decades; they’re all mini mills and/or direct reduction furnaces and yeah, the latter use natural gas. But again, you could use electrolytically producers hydrogen. There’s no reason to use steam reformed hydrogen except it’s currently cheaper.
More generally, Systems Dynamics folks call these kinds of simulation "management flight simulators". The underlying simulation approach is distinct from agent-base modelling but still useful for constructing mental models.
Problem is all such games are subject to their assumptions. Thus whatever you come up with, either it confirms my bias or I can attack it because of some obscure thing you didn't account for. There are enough obscure things for me to keep adding that you will never win (except by confirming my bias at which time someone else will jump in).
> Problem is all such games are subject to their assumptions.
Including SimCity and Cities Skylines. See:
"...through a survey of gameplay features and online discussions, I argue that CBGs present a biased urban imaginary whose underlying rules and assumptions often run contrary to contemporary best practices in urban planning and policy. Working within these constraints however, players wield considerable power to sculpt their own unique urban visions, which come to embody the ideologies of both player and developer."
In particular around automobile-centricity:
"The illusion that additional capacity solves all transit ills is compounded by the fact that no CBG to date has accurately portrayed the resultant demand for parking spaces. In C: S, vehicles may park parallel along streets and in surface parking lots adjacent to buildings. These surface parking lots are unrealistically small, however, with 10+ story buildings often having no more than a half-dozen (and often no) parking spaces. While it isn’t uncommon for dense urban environments to feature subsurface parking garages, it takes a substantial leap of faith to assume that every building in the game is equipped in this way. So what if a CBG dared to adopt a more realistic parking model? Steve Librande, a lead designer on SimCity 2013, is said to have remarked, “so much of the screen would be dedicated to asphalt that it’d be too boring to play”"
> While it isn’t uncommon for dense urban environments to feature subsurface parking garages, it takes a substantial leap of faith to assume that every building in the game is equipped in this way. So what if a CBG dared to adopt a more realistic parking model? Steve Librande, a lead designer on SimCity 2013, is said to have remarked, “so much of the screen would be dedicated to asphalt that it’d be too boring to play”"
I mean, it's also not uncommon to have surface parking garages. You can have 15 times the parking in the same land area that way, and that's what people do even in very thinly populated cities. Or suburban malls. It's not a choice between "all of the parking is underground" and "all of the parking is on the ground".
Sure, but the point is that none of these games make you account for parking or make it part of your urban planning. You don't have to build municipal parking structures or decide whether to charge for parking, or set policies around parking minimums in different classes of buildings. You don't have to deal with the blight of empty parking lots, or the challenges of reclaiming that land. There's no such thing as a park-n-ride in C:S—if you build a rapid transit line, you don't get to realize the growth associated with the greater density that would unlock in the real world because you were never bottlenecked on CBD parking availability the way the real world is.
Parking is just taken for granted as plentiful and invisible, which is why SimCity and C:S let you make a city that looks like NYC or Barcelona, but functions under the hood like San Jose.
This is problematic because non-car tranportation modes have enough trouble competing in the real world where these costs _are_ paid. How much harder is it in a fantasy where they aren't?
They are subject to their assumptions, but they also make those assumptions much more concrete. That makes them easier to version, critique, test and improve. The alternative is unstated assumptions minus the benefits, so what's to lose?
The Gathering Storm expansion in Civilization 6 adds c02 to your factory emissions, which has to be managed to avert climate catastrophes. Would highly recommend!
I was going to make note of this. The terraforming simulations of trying to make Mars and Venus livable are also interesting from this perspective. Mars is actually achievable by creating the greenhouse effect with CO2 generators, but I could never get anywhere with Venus.
I'm still baffled this hasn't been re-issued in a modern form or at least copied by someone else or inspiring another game.
Would love to make my trichordates sentient again!
There are quite a few great older games that don't have modern counterparts. Several of the Sim series of games are among them. I haven't seen a good replacement for Shadow President yet—there was one similar game called Superpower in the early 2000s that was better in some ways, but a lot worse in others, and even that's practically ancient these days. Hunter Hunted. Return Fire. Dominus.
Of course (see above comments about modern project bloat) the art team for a modern version would probably exceed the count of everyone involved in the original, there'd probably be 3-5x as many developers even if the gameplay and physics and such weren't more complicated, and so on. The credits roll would take up more disk space than the original game did.
Wasn't me that downvoted, but I'm almost positive it was sentient and not sapient as used in the game. And then I went and skimmed through a playthrough and sure enough, it's sapient, lol!
>Many things can be learnt from simulating real world systems.
This reminded me of the book "Business Engineering with Object Technology" by David A Taylor. In one chapter he goes about identifying "objects" in a manufacturing business and goes down to the level of modeling employees and the machines that they'd be using on the production line. I've never found the need to model things that granular, but it's definitely an interesting concept of "reifying" _everything_.
The reviews for that game were amazing just because of the lack of self-awareness:
[Keith Schlesinger] criticized the "not well developed" victory conditions, reporting that he got a high score in one game by "sacrific[ing] nearly all of the big and small business firms in the country to achieve the win! What kind of a victory is that?" Schlesinger found that adopting an extreme libertarian ideology including ending Medicare and many other government services resulted in almost hands-off victory although "the entire private insurance industry goes bankrupt in the first three months.
Maybe I'm alone in this, but for reasons I can't figure out, the "city simulation" flavor of sandbox games always leave me with this odd feeling of empty, like I spent a whole lot of time but ultimately wasted it and have nothing really to show for it. I can play Kerbal Space Program or Minecraft for hours and come out of it feeling like I really created something and enjoyed it. But after a few hours of SimCity or Cities: Skylines, I get this odd "wow, nothing is really accomplished here" feeling. I can't really explain it. I've tried taking breaks from the games and coming back to them, and it's always the same feeling: Well, it's 3AM and all I have is some same-looking buildings on nameless residential, commercial, industrial zones, unfinished streets ending in the middle of a field somewhere and a half-working railroad...
If I had to guess based on my similar experience, it's because the real meat of what gets built - the buildings and population - happens without direct interaction.
Factorio and Minecraft both let you take direct control over every buildable thing, but city sims make you just lay down simplified infrastructure and...wait.
At least that's how it is for me. Especially in a game like Cities:Skylines where getting more stuff to do is gated behind population counts, which you can only increase by sitting around.
I experience this as well. I find the game very enjoyable for a few hours. Unfortunately the “just one more turn” effect causes me to play SimCoty, Tropico, etc in long sessions. Towards the end of the session it just feels like I’m working on something my boss ordered me to do. I sense that I’ve spent the last six hours working, but for no money and no reward.
I bet if I played these games for an hour at a time this sensation would not occur.
How much are you into visual design and creation? I suppose Minecraft has that appeal, but it also a survival/action/crafting component not found in a city sim. For example, I'm personally confused about the appeal of Animal Crossing as a game, but from what I can see from those who gush over it, a lot of the appeal is getting to decorate and personalize an island.
Semi-related: I'm working on agent-based simulations for coronavirus, exploring our individual actions affect outcomes.
In the first simulation, I take a look at single/dual shopper households and explore how choice affects viral spread on a societal level, and then on am individual household level.
You're making me want to make a simulation to explore how different behaviors among the populace of a geographic area can influence the spread of infection. It'd be cool to be able to set policies in a region as a behavioral control, and to simulate processes like hearsay and news reading as a mechanism for the spread of accurate and inaccurate information. You could have mechanics like testing materials, ppe availability, laboratory capacity, patient turnaround, symptom severity, et. al. control the effectiveness of different policies and other processes.
Yeah that would be incredibly useful for testing policy in a simulated environment before releasing into the wild :)
One of the hardest challenges is "How do we choose realistic parameters for how the virus spreads? Infection duration? Mortality?"
But if we can incorporate the latest research and Monte Carlo simulate with ranges for the above parameters and find policies that are robust against those parameters, I think it would be an enormously powerful policy tool. Additionally, I think it could bring transparency into the decision process, something the public lacks.
If you want to contribute to the effort I started, the simulation code is open source at https://github.com/jinpan/covid-simulations. The engine is in Rust, which is then compiled to wasm.
I'd love to see more work on real-world city simulations that allow normal people to play around with them as well. Most realistic simulations right now are only accessible to experts (and even those don't tend to be at a scale to try to simulate an entire city or more). Anyone aware of others?
Sorry to double post, but I'm amazed that I seem to be the only one here who knows of this gem.
I mean, just look into the level of detail he puts into his writeup on the pathfinding algos [1]. SC4 is a whole region simulator with multi-city economics in play. Runs in Linux on Wine. Best damned $20 I ever spent.
It's fairly early on, but it's a multiplayer SimCity like game. One of the express goals is to simulate individual people/families on up to markets and such. I think the project's ideals will resonate with you; the developer wanted a game where people could cooperate to manage a complex system, and learn about system dynamics through play, which is (I presume) the role SimCity played in their life.
I'm working on https://abstreet.org, just for the transportation aspect of city sim. My goal is to make something in between VisSim (proprietary, hard to use, but industry standard) and Skylines (not based on real places or meant to model trade-offs like parking spaces and lane usage). I gave a long talk recently to the Seattle Rust meetup if you want lots of gory details: https://www.youtube.com/watch?v=chYd5I-5oyc
I don't know how often I've looked at combinatorics and yet I still harbor a secret belief that you can pre-calculate a hell of a lot more than we do and just use the result when the time comes.
Its like part of my brain lives in an alternate universe where some entire field of dynamic programming has flourished that we don't (or categorically couldn't) have here.
When I think of the fan spinning in Skylines I am transported to a memory of staring at an empty chunk of ground, watching the trees blow in the wind, and wondering if they really needed to be calculating this just right now.
There are tricks you can do with animation cycles where if all cycles are relatively prime, then the pattern only repeats every few minutes, hours, centuries. It's part of why we have 13 and 17 year cicadas. They align on a frequency of over 2 centuries, and they have near misses (appearing in successive years) around every 50 years (with a conspicuous 118 year respite in the middle that has me wondering about how public policy and cicadas work...)
> There are tricks you can do with animation cycles where if all cycles are relatively prime, then the pattern only repeats every few minutes, hours, centuries. It's part of why we have 13 and 17 year cicadas. They align on a frequency of over 2 centuries, and they have near misses (appearing in successive years) around every 50 years
You didn't explain how this observation fit into the rest of your comment, but if you're saying that putting two things on simple cycles is good enough to look acyclic to players, I don't think that's true. If you have an oak tree and a fir tree next to each other, animated in 13 second and 17 second cycles, and it takes four full minutes for the oak+fir system to repeat itself... people will still notice that the oak tree is on a 13 second cycle, and the fir tree is on a 17-second cycle. There's no rule that says you can only consider things in the context of everything else that might happen to be nearby.
I don't think it's the animations in cities: skylines that makes it a CPU-hog, I think it's the actual simulation of the various agents pathfinding, etc.
I highly recommend the factorio dev blog for anyone interested in game development - they do some amazing stuff to performantly simulate millions of discrete particles simultaneously. My 10 year old macbook can run ~10k autonomous robots before I start seeing any slowdown.
I second the motion. I would love for more realistic simulation to make its way into games. I've often thought that if my current petro-modeling job doesn't work out, I should take my domain knowledge and plug it into an educational/entertaining refinery sim game.
I keep wondering if there's a business simulation game, or a gamey simulation, that requires me to deal with the actual business side of things, instead of placing roads and snack stands. Like, to negotiate supply and demand (for b2b), build out relations in support of products, pay attention to marketing costs and otherwise flex the MBA and economic-theory muscles.
Because, dunno about SimCity, but Theme Hospital, RCT and TTD prepared me for running a business about as much as Q3 prepared me for commanding a SWAT team. And modern ‘business simulators’ that I see around look like shovelware variations on those three.
That seems like one of those things that by their nature are extremely unlikely to come about, seeing as to have a decent "win" condition you'd have to be able to understand what makes a winning business in real life.
Either no one wants to out their trade secret sauce; or there really isn't anyone that knows how.
Now I bet you could maybe stealth it by pitching a business "failure" simulator. Even then Though you may run into the same issues.
I'd also play the hell out of it. No MBA, but I absolutely love unintuitive optimization problems, and playing with hypotheticals.
I just have the feeling that if we knew enough to convincingly gamify it, it may not necessarily be a game worth playing when you could be doing the real thing instead.
I'm laughing at the mental image of contests wherein people bring their pet company in for evaluation for entry into a business program or something.
Well conditions also aren't homogeneous, uniform, or exhaustively known which makes tbings better and worse.
For a real life vs simulation example: there isn't much practical use in knowing how to be successful in building the first rail lines for instance - they have been there for centuries now the conditions aren't the same at all.
I went through Capsim (https://www.capsim.com/) as a part of my MBA program. It covers the simulation of R&D, Marketing, Production, and Finance - not so much Economics.
I remember playing a game that for once simulated "business"
It was called "Executive Suite" by the amusingly named "Armonk Corporation" :)
"Executive Suite is a text-based business social simulation game" ... "The game is a prototypically self-aware humor-driven (yet with serious aspects)"
As a new engineer, you are subject to the whims of
unorganized project managers.
A last minute assignment conflicts with a special
seminar MMC has enrolled you in.
Your project manager, Harriet Nance, tells you to
handle the situation.
What do you do?
1 - Delay the work
2 - Skip the seminar
3 - Take the work to the seminar
Press the number of your choice and press the Enter Key:
Capitalism Plus is much more like that. It's still very, very idealized, but you are spending your time mostly managing supply lines and marketing over the physical placement of things...
SimCity is what got me interested in the IT decades ago and it's one of the two games I have ever bought (the second being Civilization) and played (not counting Tetris :) ). It will always have a special value for me.
I have a similar starting point. Civilization was one of the first the only games my computer could play at the time. I remember racing in the middle school in the mornings to get some time on SimCity 2000 with my buddies. We had one city going for so long...
Great read. I absolutely loved this story. I was a teenager when I first played Sim City and remember fondly how amazed I was back then at the possibilities for the future then. It certainly felt like adding a continuation that this story added to. We need to nurture more people like John Hiles and make sure they don't get devoured by narcissistic psychopaths in business suits. Again, thanks for posting this, I enjoyed very much listening to this story.
>You were meant to destroy everything once in a while.
>This was even the main takeaway for some people. [...] SimCity 2000 [...]
Nooo! I didn't like the disasters in SC2000. If your city becomes too bad a delegation of citizens gently escorts you from your mayor's office[1] (game over). Also, getting disasters under control takes work and there is no progress in the city during disasters.
If there is a fire, I had to place firefighters around the fire. Sometimes it made sense to take the bulldozer and remove structures, which would have been fuel for the fire. I hope people didn't miss their houses :^). After I had successfully learned how to always get the different disasters under control, I disabled them in a menu.
BTW I wish a Windows emulator with SimCity2000 for Windows would be sold. Right now only the DOS version is sold.
The fun was stacking earthquakes and volcanos. Simulated meteor strike! Or mashing "hurricane" like 10x in a row followed by a few tornados to give you a superstorm that'd flood most of the map and leave 80% of your city leveled. Once a city got to a point I didn't really care to build more, recovering from superdisasters was where the fun was at (for me).
I can't find any business simulation companies that do graphics the way Maxis/MBS did. It seems like most business simulations today are dashboard-driven: for example, Cesim offers a retail store simulation where you move the unit price toggle and then the sales volume meter goes down while the profit meter goes up[1] but there's no "SimStore"-type visual model where you can see agent-based shoppers bopping around while filling their carts.
Are there any business sims that offer graphical models? (Don't those seem more fun -- and thus more effective -- for employees to play and learn?)
SimEarth really taught me the importance of critical mass. You could peter out if you didn’t marshal your resources to reach the next stage of the game.
Oxygen Not Included is my simulation game of choice atm. It really punishes you if you don’t realize how quickly you’ll deplete some finite resource.
Interestingly, SimTower was not developed by Maxis but by the Japanese OPeNBooK, and was originally released in Japan as "The Tower". Maxis then published it in US as part of the Sim series.
OPenBooK went on to develop Yoot Tower ("The Tower II" in Japan), a sequel to the original game, in 1998.
> Jeff Braun was the only manager for a staff of 32 people – even for the financial staff, a problem that was swiftly corrected after Maxis got audited by the IRS.
Why does the IRS care how one organizes their company?
That might imply incompetence on their financial staff's part and his inability to handle it properly. That or maybe dumb local tax codes like based on job titles.
> The game was inspired by research on real-world urban planning concepts
This is not the anecdote I'm familiar with. My understanding was that Will (or someone else) noticed that people were having a little too much fun farting around in a content creation tool they had written in-house and they decided to make it into a game of its own.
The concept for SimCity was inspired by how much fun Wright had playing around with the map design tool for Raid on Bungeling Bay. The design of SimCity was inspired by Wright's research on urban planning and system dynamics during its development.
I think I am using a different definition of "inspired" than the author.
Your inspiration for the love song you just wrote is the special someone you have your eye on. The lyrics are informed by every other case of New Relationship Energy you have ever had, with bits borrowed from friends, books, and movies.
Was Will really studying urban planning for giggles, put 2 and 2 together and got 5? Or was he working on a hunch about a city planning game and what he found cemented his hunch?
Here are some talks about Modelica: https://www.youtube.com/watch?v=-mvEUuc-sWE&t=153 and https://www.youtube.com/watch?v=39F___xyI0k
Here is a complete course if you want to dig in: https://spoken-tutorial.org/tutorial-search/?search_foss=Ope...
And if you're wondering "Will it emscripten and run in the browser?", the answer is yes, of course: http://tshort.github.io/mdpad/mdpad.html?Modelica.Electrical...
Context: I'm researching a new book on ordinary differential equations through various real-world simulations, so like an ODEs theory book, but actually applying the concepts to real world systems. Get in touch by email if you're interested in seeing advanced preview of some of the projects that will be developed this summer. Contact info in profile.