iMSISS Labs

In Management Science we were doing Decision Analysis, specifically Decision Trees. Decision Trees is a way of modelling decisions and states of nature to compute the "most efficient "best" choice based on probabilities of states of nature. I've built a simple Decision Tree renderer for a custom language used to describe the trees. I may develop it further so that it actually calculates best paths, but for now it simply draws them.

The graphs are fairly ugly but that's a limitation of the graphics library really, and I don't want to spent too much time adapting it to a better one.

Generate Your Own

After reading through the examples try generating your own decision trees:

The Language

[Node Name] Square Brackets Represent Decision Nodes
(Node Name) Paranthesis represent States-of-Nature Nodes
Angle brackets represent Leaf nodes
-> Arrows represent sub-nodes
{ Node, Node, Node } Braces represent a comma-separated list of nodes

Examples

[Buy Or Sell]->{ <Buy>, <Sell> } Produces:

 [Company Strategy]->{ 
			(Sell)->{
				<Strong Market = 500,000 eur>, 
				<Weak Market = 320,000 eur> 
			}, 
			(Buy Competitor)->{ 
				<Strong Market = -400,000 eur>, 
				<Weak Market = -120,000 eur> 
			} 
		}

Produces:

 [What To Do]->{ 
				(Expand)->{
					[Strong]->{
						(Build At Home)->{ [Land Is Good]->{ <Build>, <Do Not Build> }, [Land Is Bad]->{ <Build>, <Do Not Build> } }, 
						(Build Abroad)->{<Get Planning>,<No Planning>}
						} ,
					<Weak>
					}, 
				(Keep)->{
					<Strong>,
					[Weak]->{ [Sell Business]->{ <Sell Abroad>, <Sell At Home> }, <Get Loan> }
					}, 
				(Sell)->{
					<Strong>,
					<Weak>
					} 
			}

Produces: