If I have understood correctly, your design is to keep the cost of mining your currency stable by disincentivising improved technology which can mine more profitably.
This is part of it, but not all. A lot of energy has to be "burned" by the miners to even begin to create some new currency, meaning that hopefully the currency must be valued somewhat above its cost to produce before mining can even begin. In addition to this, so that lots of energy must not be wasted in creating money, and to incentivize using the money as money rather than a speculative vehicle, lots of free money is given away when people do mine so that the currency becomes unprofitable to mine again relatively quickly. If the currency is below its cost to produce, no one will likely produce it, and thus incentivizing people to accept it on the hope of it returning again to its higher value. This should be a fairly powerful motivator considering how much faith is really required for any currency to be successful.
This may keep the currency more valuable than others to mine, but I'm not sure this will stabilise the value of the currency.
Neither am I, but I think it is the most effective design possible given what we know about how cryptocurrency will be used, which is almost nil at this point.
I am working more on a detailed design and a basic simulation to help develop an algorithm which can lead to the network regulating its own rate of currency circulation.
Basing it on transaction activity won't work well, especially if you raise tx fees in response to increased activity. There will be many other ways to transact the currency that are "off-chain" and won't be monitored by the network, throwing off the calculations. This will in fact be encouraged if you raise tx fees and will further skew the data.
I considered using an idea where the difference in the
amount of transactions could be monitored to potentially gauge inflation or deflation, but this really can't account for the changes in how cryptocurrency will be used in the future, and thus will likely be wrong more often than it is right. Maybe after the world has been using cryptocurrency for a hundred years or so and we have gobs of real data, something could be devised that will be very successful, but at this point it is just a guessing game, and I'm fairly sure every guess is going to be wrong. There are just too many factors to account for, and not nearly enough information to draw on.
