Any sustainable group of interacting agents have to make decisions on how to work together to achieve the best outcome for the group. Human groups use a variety of strategies including markets, representative governance, or assigning authority to some members as ways to work together.

In small groups, such as families, decisions tend to be by consensus. Larger groups find agreement challenging and so we invent strategies to overcome the inevitable disagreements. The simplest approach is to abandon consensus and to use centralised enforcement. Unfortunately, the cost of centralised decision making with enforcement increases exponentially with the number of members in a group.

One way to reduce costs is to form subgroups and get agreement between the subgroups. The subgroups are stable and in turn, we can develop a consensus between the subgroups.  With this approach, costs increase linearly with the number of members. One method is to create subgroups around each entity. An organisation of many entities is one such group. A tribe is another. Individuals have subgroups with whom they interact for given purposes. In general, any entity interacts with many other entities and each purpose becomes a subgroup.

Entities do not have to be people or organisations. An entity can be a loan and we can create very large systems with loans where all the parties both borrowers and lenders to the same loan form a subgroup.

The Internet provides us with a new set of tools to reduce the cost of consensus because we can create many more subgroups for all types of entities and handle the interactions between them efficiently. It enables the group of entities with which an individual interact for a purpose, to be a subgroup. This grouping of interacting objects is often called a swarm and the behaviour called swarm intelligence. With the Internet, we can create many subgroups and control them so that leakage of data outside the subgroups is minimal. For human systems, this leads to enhanced privacy and freedom from surveillance.

An issue with consensus groups is what happens when a person or entity will not agree or breaks the consensus. Consensus on enforcement is low-cost with small groups.  Peer group pressure is an example.  When someone does not conform, they are left out of the group but can join another group.  When a merchant’s prices become too high, you leave the group of their customers and join another merchant’s group. The more variety, the more alternatives, the greater the chances of consensus.

This post outlines some examples of how the Internet enhances agreement by having entity based groups for different purposes.

  • Where is it?
  • What is my address?
  • Who am I?
  • What price should I charge?
  • How can I find a place to live?

Where is it?

Here the objective is to discover the physical location of an object.

At some point in time let us assume that all objects know where they are and they know the distance from every connected object. When an object moves, the object that moved is the only one that has to recalculate its position.

To answer the question, “Where is it?” the swarm including the object, knows where they are at all times if they know the distance to connected objects. The algorithm to calculate positions is the only thing that needs to know the position of each object and it only needs to know while it is calculating. It asks each object, in turn, where they are, does its calculations and reports back any change in position to each object. The algorithm does not remember the position of any object. Objects never know the location of other objects unless the others wish them to know.

To find the distance between objects each object keeps a record of previous measures of distance and can use those records to check and calibrate current measures of distance.

If an object wants to exclude another from its group it can do so by refusing to tell the other where it is.

The difference between this and current methods of location is that an object is the only one who knows where it is at all time. Its location is a private attribute that it provides to the known algorithm.  It only provides its location to others who it approves and only via known algorithms.  If a group member goes against the wishes of an object and provides the location without permission then, the object stops communicating with the offending member and tells others of its concerns.

What is my address?

A common utility is the directory problem.  You wish to make a connection with someone, but you do not know where they are. The traditional solution is to have a centralised directory where you put your address, and people ask the directory.  The distributed solution is for you and those with whom you wish to communicate only to know your address.  That is, you have a directory of your own.  Someone outside your group knows who you are but does not know how to connect to you. They ask their connections do they know your address. If they do not know then they can ask their connections and so on.  You can set up rules on whether you wish to connect by being asked and you only connect to those who agree.  Social networking sites use this approach. The difference in a distributed system is that there is no need to have a social networking site.  The algorithm is the social network.

Who am I?

The objective is for your electronic identity to remain consistent. It means when two parties reconnect they are confident they know each other. At some point in time let us assume that all objects know their identities and they know how to identify themselves to other objects to which they are connected. Let the location be part of a measure of trust in the identity of the other party. If there is a change in trust, such as a person is now in a different physical location, then the swarm can ask all connected parties to recalculate their measure of trust and see if they are prepared to continue to accept the identity of the changed party.

What price should we charge?

The traditional way of setting prices is through demand markets. When demand is high, the price goes up. When demand is low, the price goes down. Setting prices this way works well when there are one-off sales. It is expensive when the sales are commodities and buyers make repeat purchases. If fails when there are monopoly buyers or sellers.

One way for consensus pricing is to agree on a base price for all production. Prices can increase to some maximum but rather than all the growth in income going to producers some goes to those who restrain their consumption. The increase could be in the form of Rewards for non-consumption. The base price, the Rewards and rules of operation use consensus to come to an agreement on prices.

How can I find a place to live?

In our society, we agree that it is desirable for most people to have a permanent place to live. We know that people like to have space in which they are comfortable, safe and private. We have developed ways to achieve this. The most common way in industrialised societies is through home ownership.

Some people are unable to obtain a home of their own through the usual channels. Consensus Homes establishes a way for members of a group to each have all the rights of homeownership to a dwelling.  They obtain the rights by consensus of the group. Each person wishing to participate negotiates with others in the group and decides to live in a particular dwelling. They live with or without others look after it and agree to buy the mortgage over the property through regular payments.  The person builds equity in the mortgage that they can transfer as a deposit for a traditional home purchase or to another dwelling within this or another group.  Enforcement comes by people who fail to pay regular payments having to leave the dwelling.

One thought on “Making Decisions by Consensus

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s