Abstract
Many topics related to stability of equilibrium, equilibrium selection, transitional dynamics, the long-run evolutionary dynamics of economic processes or other microeconomic or macroeconomic concepts are of the most importance in economic studies and often involve a set of optimization techniques able to provide the best solution under specific conditions. One of the main issues that should be studied as an optimization algorithm is applied is the total time required for that process to get the optimum. Even in this era of the New Technologies, as computers are designed to work faster and faster, it is welcomed to know to which extent a specific function is welcomed to be optimized by using a specific algorithm, or to which extent the time required by that algorithm to get the optimum is a polynomial or an exponential one. Among the classical methods that we can use to drastically reduce the dimension of the transition matrix P of the Markov chain attached to (1+1)-EA, the comassability of the states is a very important one. For some fitness functions this comassability is possible, but for other it is not. This paper aims at making an exact description of certain unimodal functions which lead to a transition matrix P comassable in relation with a partition of the states space.