Has the AI function improved for TS 2019?

AI does not have enough to solve for "X" or the way out given logical solution?

Considering all the interesting types of Robots I read about, whose programs are driven by pure AI with some salting of Human feelings, emotion, morals, I would think that our AI has a very bright future of which we will not even recognize him from today in the near future of what it will be able to do. Just ask Elon Musk or Jeff Bezos Etc........

:( Currently Our AI is still a young baby, with not enough knowledge to work out of difficult situations, IE the Yellow Light and proper adjustment with Trainz Speed.......Instead of dropping speed immediately to an unreasonable level.....He should be able to adjust his speed properly way in advance with his Arithmetic and Formulation sensory input!!!!!

It would seem that simple in Theory, however it doesn't occur! Especially when we know he can sense electronically seeing all the Tracks ahead, signals a long ways down the Track? :hehe:

Which brings me to another question, if our Software is written to calculate Trainz logic, such as like Chess moves against an opponent to gain a given result, IE WIN the Game, then why can't it formulate the proper paths and moves thru switch puzzles, oncoming trains etc..........

I was taught a long time ago that in an algebraic expression, both sides need to agree, or the formula needs to be written correctly and you can solve for unknown "X" factor etc. or ?

https://sciencing.com/explain-basic-prealgebra-equations-5928250.html

I am no math major by a long shot, but I do know, down to the core, our computers run in 1's and 0's the bit is either on or off, electrically speaking, so something is really off here, if this AI cannot get thru the conditional logic needed to solve "X" or the proper path and speed to move our trains, against, around or stop (variable of Change condition) getting us from point A to B or C Etc.....Then a piece of the fuzzy logic is missing from the equation...........maybe a few pieces of "X" in theory.....

Conditional probability is the in probability theory, conditional probability is a measure of the probability of an event given that another event has occurred. If the event of interest is A and the event B is known or assumed to have occurred, "the conditional probability of A given B", or "the probability of A under the condition B", is usually written as P, or sometimes P or P.Wikipedia

We/I await for the true Power in AI driven environment to rise above in consistency, path solving etc......For now the jury appears to be still out on that one for now.........:sleep:
 
Actually the AI are following the NORAC Rules.

https://en.wikipedia.org/wiki/Northeast_Operating_Rules_Advisory_Committee

[FONT=Verdana,Arial,Tahoma,Calibri,Geneva,sans-serif]http://thebecketts.com/images/NORAC 8th Edition NJT.pdf

[/FONT]There are many, many rule books and operating procedures, but Auran at the time chose the rules from the NORAC rulebook to govern the AI operations. According to the rules, a yellow signal for example means the AI must operate at medium or half-the designated speed. A flashing yellow indicates that the following signal is a solid yellow, and to drop to medium speed and prepare to stop at the next red signal.

The issue here is the AI is just better at not only seeing the signals before we do, but also can react a lot faster than we can.

I recommend saving the PDF noted above it's worth looking through and it makes the AI operations more logical than we think.
 
Hi Blue

I don't know if you are aware of the conditional driver commands created by "trev999" on the DLS that can use variables to allow consists to make decisions about where they are going or that can use variables to allow industries to call for deliveries or collections when the industry needs them and not when you decide to send them.

The "Input Table" rule (built in) is used to set up variables and the conditions that they can use eg set, increment, decrement, greater than, less than, random etc.

By using a random number which can be anything between 1 and whatever number you want (within reason) you can use a basic conditional probability for decision making by a consist. As an example I have a modified version of the Clovis Sub route by "dermmy". A train approaching Belen Yard from either California or Albuquerque has 4 possible destinations locally. They can go for refueling, they can go into the main yard, They can bypass the yard and head for El Paso or they can bypass the yard and head East on the Clovis Sub. By generating a random number you can allow each consist to decide its own destination based on probability tailored for the type of consist.

By using variables I only need to create one session which is very unlikely to ever run the same twice.

I also use variables to control single line sections of track and it is perfectly possible by using 2 variables to have a method of allowing trains traveling in the same direction to follow one another into a section while preventing any oncoming consist from entering until all of these consists have cleared the section.

Of course all of this takes a little time and effort to get to grips with but by using such methods as these, combined with Enhanced Interlocking Towers, it is quite possible to bend the AI to your will and create realistic railway operations with the AI.

Regards

Brian
 
Back
Top