cdemers / go-randomwalker

Package randomwalker provides a parametric random walk generator.

Simple Random Walker

Simple random walker to be used for non cryptographic simulations, like natural phenomenon models or finance

Package randomwalker provides a parametric random walk generator. You can instantiate a random walker that will start at a specific point, that will stay between an upper and lower boundary, that will vary on each step by a specified maximum magnitude.

Also, if you expect to use this walker for financial models, this walker is going to give you something close to a Gaussian random walk, and not a Brownian random walk.

Calling NewRandomWalker() returns a Gaussian like random walk generator. Once created, you can call Step() on this object to get the first and then the subsequent random walk value.

It's signature is NewRandomWalker(origin, min, max, maxDynPcnt float32).

• The origin parameter sets the initial walk value, it's starting point.
• The min parameter sets a limit on how low can the walk go.
• The max parameter sets a limit on how high the walk can go.
• The maxDynPcnt parameter sets the limit on how much can a walk step vary from it's previous value. Setting this value at zero will prevent the walk from varying at all, leaving it constantly at it's initial value. Setting this value at 1.0 will allow the walk step to vary from -50% to +50% of it's previous value (the difference between -50% and +50% totaling 1.0). And setting this value to more than 1.0 is permitted, for example setting it to 4.0 will allow a step to vary between -200% and +200% of it's previous value.

Calling Step() afterward will return a float32 value.

Using NewRandomWalkerWithRandSource() also returns a random walk generator but that here you can provide your own random source object if needed.

Example

var origin, minimumWalkValue, maximumWalkValue, maximumVariationPercentage float32

origin = 50
minimumWalkValue = -25
maximumWalkValue = 75
maximumVariationPercentage = 0.01

rw = NewRandomWalker(origin, minimumWalkValue, maximumWalkValue, maximumVariationPercentage)
value = rw.Step()
// _value_ will be somewhere between (50 + 1%) and (50 - 1%).
value = rw.Step()
// _value_ will now be it's previous value plus or minus 1%.

go-randomwalker's Issues

Thread safety

go-randomwalker/randomwalker.go

Line 64 in decad42

rw.current += (rw.random.Float32()*2 - 1) * maxDynamic

Maybe it should be wrapped in a sync.Mutex Lock()/Unlock()