Page 178 - Compact Numerical Methods For Computers
P. 178

One-dimensional problems                    167
                      simulate the operation of the scheme. The results are given in table 13.2. The
                      arrests are drawn from a uniform distribution on (0,400).

                                  TABLE 13.2. Simulated operation of an income insurance program.
                      Premium =       92·90                 86·07                109·92

                       Period  n j       f i (P)     n j       f i (P)      n j      f i (P)
                         1      0      193237·50     17       158630·94    188       3029·50
                         2      2       399533·94    232       71952·31    315     146098·69
                         3    279       289934·12    317     -123660·62    194    -172183·94
                         4    124       357566·31     67      -43578·75    313    -344982·00
                         5    374       130609·06     74       40115·38     35    -210099·75
                         6    356      -92904·75      55      156355·50      7      -21385·19
                         7    101      -34802·94     152      165494·81    127      51636·50
                         8    281      -183985·87    3 0 4     -7034·69    387     -180003·12
                         9     2 3      -49546·25    1 1 3    35341·00      55      -44374·06
                         10   117          -0·69     1 8 1        -0·81    148         -0·69
                       Total  1657                  1512                  1769
                      Function evaluations
                        to find root        10                    14                    11
                      (Total benefits)/(number
                        of premiums paid)   99·42                 90·72                106·14
                      The last entry in each column is an approximation based on no interest paid or earned in the fund
                      management. Thus
                                          approximate premium = total arrests * b/(n* T)
                                                         = total arrests * 0·06.
                      These examples were run in FORTRANon an IBM 370/168.
   173   174   175   176   177   178   179   180   181   182   183