Please enable JavaScript.

Coggle requires JavaScript to display documents.

ESA Week9 - Coggle Diagram

- - - - note that colour and size can be used for continuous but symbol cannot. use symbols only for discrete
      - if you want, you can convert 1 and 0 to as.character(...) so that it appears as "1" and "0" with only 2 separate colours in the graph
    - - paretosorting_adv(input, "MAX"/"MIN")
        the input must be a matrix
      - the paretosorting_adv function can process 2,3, or more objectives. when you have 2 objectives, you might have 16 dominated solutions, but when you have 3 , you might only have 4 dominated solutions
      - parallel coordinates, ideal solution is a line at the top of the plot for maximisation functions
        
        groupColumn: you can have multiple colours using this variable
        
        columns:
        
        scale = "uniminmax" from 0 to 1
      - radar plots good for a few solutions
        
        remember to use rbind to specify min and max
        rbind(c(1,0,1),c(0,-1,0),mydata_)
        
        index matrix by [row,column]
        [row1:row2, column]
        
        to index multiple rows of the matrix, you can use
        c(matrix[8,1],matrix[25,1])
        
        note, radar plot, usually we use the normalised dataset
        
        rbind concatenates vertically, by rows
      - use which.max/which.min to find the index
      - implementation of the utopian point method
        
        identify pareto efficient solutions and store their indexes in a vector
        
        find coordinates of utopian and nadir objective vector : concatenate the max / min values of all the objectives
        
        normalise the data and check the utopian point is indeed 1,1,1 for max and 0,0,0 for min
        
        create a vector of distances and find the min using a for loop
        
        find the min, as well as its index among the vector containing the pareto optimal solutions
        
        find the index of the min among the entire dataset --> use the index[which.min(Distance)] function as the index and Distance vectors are of the same length. The index vector contains the indexes of the pareto efficient solutions among the entire dataset, whereas the Distance vector lists the normalised distances of the Pareto efficient solutions
        
        plotting in a 2d scatter plot
        
        need to create another data frame, identify the point containing the selected best solution, the Utopian point, as well as the Pareto optimal solutions
- - - - distance between two points. take the square root of (the difference in y coordinates + the difference in x coordinates)
      - use normalisation factor to avoid objective functions that have a range of variabiltiy and different units of measurements
        
        diff formulas for min and max formulas