Search device

Abstract

This search device 1 is provided with: a backcasting unit 11 that calculates, for each point in time, allowable ranges for both natural environment and socioeconomic conditions that are permitted in order to reach an ideal future state, by going back in time from the future; and a forecasting unit 12 that calculates, by using natural environment- and socioeconomy-related parameters, both natural environment and socioeconomic conditions in time series from the present to the future as human activity paths, and outputs, as allowable human activity paths, the portions of the human activity paths that fall within the allowable ranges.

Description

Search device

[0001] The present disclosure relates to a search device.

[0002] Human activities place a burden on nature, and the impact of this burden is now manifesting itself in the form of climate change. Therefore, quantitative scenario assessments are being conducted by generating narratives (narrative descriptions) through discussions among experts, defining parameters such as economic growth rates in line with the narratives, and then predicting future indicator values ​​such as gross domestic product based on these parameters (see Non-Patent Documents 1 and 2).

[0003] “3.3 Target finder”, [online], [Retrieved March 19, 2024], <URL: https: / / jgcri.github.io / gcam-doc / user-guide.html#target-finder> Kurahashi, “Explanation: Model estimation and inverse simulation method”, [online], [Retrieved June 19, 2024], <URL: https: / / www.u.tsukuba.ac.jp / ~kurahashi.setsuya.gf / doc / is2.pdf>

[0004] Previously, while forecasting indicator values ​​mainly in a forward direction, a time-series human activity pathway was calculated that would keep radiative forcing, an indicator of climate change, within a target value. However, even if human society acts in accordance with this human activity pathway, there was a problem in that it was not clear whether problems would arise in the natural environment until they actually occurred.

[0005] The present disclosure has been made in consideration of the above circumstances, and an object of the present disclosure is to provide a technology capable of searching for human activity routes that are acceptable for both the natural environment and the socio-economy.

[0006] A search device according to one embodiment of the present disclosure includes a backcasting unit that calculates, for each time period, the allowable ranges of both the natural environment and socioeconomic conditions that are acceptable for reaching an ideal future state by going back in time from the future, and a forecasting unit that calculates the time-series natural environment and socioeconomic conditions from the present to the future as a human activity path using parameters related to the natural environment and socioeconomic conditions, and outputs the portion of the human activity path that falls within the allowable ranges as an acceptable human activity path.

[0007] According to the present disclosure, it is possible to provide a technology that can search for human activity routes that are acceptable for both the natural environment and the socio-economy.

[0008] FIG. 1 is a diagram showing the functional block configuration of a search device. FIG. 2 is a diagram showing the functional block configuration of a geographical calculation unit. FIG. 3 is a diagram showing the functional block configuration of a backcast unit. FIG. 4 is a diagram showing the functional block configuration of a forecast unit. FIG. 5 is a diagram showing an image of an allowable range of a state. FIG. 6 is a diagram showing a method for extending an allowable range of a state. FIG. 7 is a diagram showing a search result of an allowable range of a state. FIG. 8 is a diagram showing a method for limiting an allowable range of a state. FIG. 9 is a diagram showing a reference example of setting a tipping point boundary value. FIG. 10 is a diagram showing an example of obtaining an allowable human activity route. FIG. 11 is a diagram showing an example of obtaining an allowable human activity route. FIG. 12 is a diagram showing a search flow for an allowable human activity route. FIG. 13 is a diagram showing the hardware configuration of a search device.

[0009] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. In the description of the drawings, the same parts are designated by the same reference numerals and the description thereof will be omitted.

[0010] This disclosure relates to the technical field of exploring, evaluating, and analyzing scenarios of human activities aimed at achieving sustainable socio-economic growth. In particular, it relates to a technology for calculating the trajectory (human activity path) that connects the future state of the natural environment and socio-economy over time as a result of human actions.

[0011] In order to solve the above problems, the present disclosure has the following features.

[0012] The first is to quantitatively evaluate what state the natural environment and socio-economy should be in between the present and the future target point, using an inverse simulation of a system that includes both the natural environment and human society, in order to achieve an ideal future state.

[0013] The second is to constrain the allowable space for reaching the ideal future state, taking into account tipping points in both the natural environment and human society.

[0014] The third is that in the real world or in forward simulations, it becomes possible to understand in advance whether the ideal state can be reached in the future, even at an intermediate stage.

[0015] [Overall Configuration of Searching Device] FIG. 1 is a diagram showing the functional block configuration of a searching device 1 according to this embodiment.

[0016] The search device 1 is a device that searches for "allowable human activity routes" based on the "current state" and "target ideal future state" of the natural environment and socio-economy, and "evaluation scenario parameters" for determining future human behavior.

[0017] To perform this search, the search device 1 is equipped with a backcasting unit 11 that performs inverse analysis (inverse simulation) of the state of the natural environment and socio-economic situation from a future starting point, a forecasting unit 12 that forecasts the state of the natural environment and socio-economic situation from a present starting point, and a global calculation unit 13 that calculates the state of the natural environment and socio-economic situation.

[0018] The backcasting unit 11 receives (inputs) the "target ideal future state," calculates the conditions (range of acceptable states) that indicate the range within which the natural environment and socio-economic states (state variables) must fall at each future point up to the target future in order to reach that ideal future state, and passes this "range of acceptable states" to the forecasting unit 12.

[0019] The forecasting unit 12 receives (inputs) the "current state" and "assessment scenario parameters," receives from the global calculation unit 13 the state of the natural environment and socio-economic situation when calculations are carried out forward in time from the "current state" using the "assessment scenario parameters," and extracts only the "range of allowable states" from the "human activity path" and outputs it as the "allowable human activity path."

[0020] The backcasting unit 11 and the forecasting unit 12 use the geographical calculation unit 13 to calculate the state transitions of the natural environment and socio-economy in the backward or forward direction of time.

[0021] [Scenario] A scenario is a scenario used in "scenario analysis."

[0022] In other words, a scenario is a process in which some assumptions are expressed in natural language in an area of ​​high uncertainty, and then the natural language expressions are converted into numerical parameters, which are then set in a mathematical calculation model to calculate future conditions. In other words, a scenario is used when you want to judge the merits of decision-making while depicting future developments.

[0023] The scenario according to this embodiment is comprised of the "human activity pathway" calculated by the forecasting unit 12, the assumed narrative (narrative description) on which the scenario is based, and the "evaluation scenario parameters" that represent the narrative as a set of quantitative parameters. The "human activity pathway" is the result of determining the narrative, designing experimental parameters in line with the narrative, and quantitatively calculating what the future world will be like with those parameters.

[0024] [Functions of the Earth Calculation Unit 13] FIG. 2 is a diagram showing the functional block configuration of the earth calculation unit 13. As shown in FIG.

[0025] The global calculation unit 13 has the function of calculating the state of the entire Earth while taking into account the interaction between the natural environment and socio-economy. A monolithic model that includes everything may be constructed and used, or existing natural environment models and socio-economy models may be coupled and used. Each model may be constructed using a data-driven, lightweight model based on machine learning, etc.

[0026] The geological calculation unit 13 calculates the values ​​of the state variables at the next time using the state variables of the natural environment and socio-economy at a given time. To perform this calculation, the geological calculation unit 13 includes a natural environment calculation unit 131 that calculates the state variables of the natural environment in the backward or forward direction, and a socio-economic calculation unit 132 that calculates the state variables of the socio-economy (economic society) in the backward or forward direction.

[0027] Interactions occur within the natural environment, within the socio-economy, and between the natural environment and the socio-economy, and each functions as input to the other. In other words, the natural environment calculation unit 131 and the socio-economy calculation unit 132 may calculate state variables for only the natural environment or only the socio-economy, or may calculate state variables that involve interactions between the natural environment and the socio-economy. In Figure 2, dashed lines represent interactions within the calculation model and interactions between calculation models.

[0028] The global calculation unit 13 receives decision-making information from the human society side as "assessment scenario parameters" and can correct each parameter of the natural environment and socio-economy. For example, on the natural environment side, it can set changes in land use due to afforestation (bare land → forest). On the socio-economy side, it can set information such as what percentage of carbon tax rate to be applied in what year, and policies for converting energy use.

[0029] [Functions of the Backcasting Unit 11] FIG. 3 is a diagram showing functional blocks of the backcasting unit 11.

[0030] The backcasting unit 11 includes an allowable range calculation unit 111 and an allowable range output unit 112 .

[0031] When a "target future ideal state" is set, the tolerance calculation unit 111 determines what range (tolerance range) the state variables of the natural environment and socio-economics must fall within at a point in time going back a predetermined time interval in order for the set future ideal state to be reached.

[0032] The allowable range represents the range (boundary condition) of state variables with a time before a future time point that can reach the ideal state at that future time point. The minimum value (earliest time) of the allowable range may be set to the current time point, or any time point between the current time point and the set future time point.

[0033] The allowable range calculation unit 111 searches for reachability by randomly scattering model parameters, which are decision-making information on the natural environment side and the socio-economic side. At this time, an upper and / or lower limit may be set for the search range based on the current performance of the natural environment, socio-economic policies, etc.

[0034] The allowable range output unit 112 converts the allowable range of the state variable for each time output from the allowable range calculation unit 111 into a format that can be used by other functional units and outputs it as a "range of allowable state."

[0035] [Functions of Forecasting Unit 12] FIG. 4 is a diagram showing functional blocks of the forecasting unit 12.

[0036] The forecasting unit 12 inputs decision-making information from the human society side into the geological calculation unit 13 as "evaluation scenario parameters" and searches for future human activity paths in the forward direction.

[0037] To perform this search, the forecasting unit 12 includes a future route calculation unit 121 , a route confirmation unit 122 , and a future route output unit 123 .

[0038] The future path calculation unit 121 uses the global calculation unit 13 to calculate the path (future path) of the future state of the natural environment and socio-economy.

[0039] The route confirmation unit 122 confirms whether the future route is included in the "range of allowable states" output from the backcast unit 11 at each future point in time included in the future route.

[0040] The future route output unit 123 outputs a "future route" ("human activity route") that falls within the "range of allowable conditions" as an "allowable human activity route."

[0041] [Function of the Tolerance Calculation Unit 111 of the Backcasting Unit 11] The tolerance calculation unit 111 calculates the tolerance at a future time T ｆ By performing a reverse analysis from the ideal state of ｆ The allowable range of the state at a predetermined time t within which the ideal state can be reached is estimated.

[0042] Inverse analysis is a method for estimating, by working backwards, what initial conditions, loads (forcings, inputs), and boundary conditions led to a given state. There are also methods for estimating the governing equations themselves (see Reference 1).

[0043] Reference 1: “What is inverse problem analysis (reverse analysis)? An approach using mathematical optimization and simulation”, NTT DATA Mathematical Systems, [online], [searched March 19, 2024], <URL: https: / / www.msiism.jp / article / solution-inverse-problem.html>.

[0044] The current time is T ｐ The lower limit of the estimated time range is T ｌ As a future time, T ｆ Estimation starts backward from the lower limit T ｌ (T ｐ ≦T ｌ <T ｆ ) and perform a back-analysis of the state tolerances until the state tolerances at

[0045] Specifically, the lower limit time T ｌ From a future time T ｆ The time up to the current time point T is divided into predetermined time intervals, and the reverse analysis is performed for each time interval. ｐ State of future time T ｆ The distribution boundary of the state of the previous time step, the future time T ｆ The distribution boundary of the state of the time step two steps before is sequentially calculated from the distribution boundary of the state of the time step two steps before. The time width may be set arbitrarily as long as it is equal to or less than the time width that is compatible with the calculation conditions of the global calculation unit 13. It may also conform to conditions that stipulate the upper limit of the time step when performing model calculations, such as the CFL (Courant Friedrichs Lewy) condition.

[0046] Hereinafter, an example of determining the allowable range of the state of the time step n+1 steps before from the allowable range of the state of the time step n steps before will be described with reference to FIGS. ｐ is set to 0, and at a future time T ｆ This refers to the state n time steps before. Hereinafter, this will be referred to as the state n steps before.

[0047] The following method is an example. Any method can be used as long as it can determine the state, boundary conditions, and load at the previous time required to reach a certain state when that state is defined.

[0048] Step 1: Tolerance R n steps before ｎ Points are set randomly for the entirety of the equation in a Monte Carlo simulation manner, and each of these points is set as a point to be reached n steps ago (a point reached from n+1 steps ago).

[0049] This point is used to search for the allowable range of n+1 steps before that can reach each point in the next step 2. The point is within the allowable range R ｎ As mentioned above, the allowable range R is set not only on the boundary but also within it, based on the current performance of the natural environment and economic and social policies. ｎ An upper and / or lower limit may be set for the

[0050] Step 2: Tolerance range R n steps before ｎ Inverse analysis is performed at each point of the following.

[0051] Step 2-1: Create the initial conditions for searching the allowable range for the n+1 step before using the distribution up to n steps before.

[0052] Figure 5 shows an image of the case where there is one evaluation index. ｎ＋１ is the initial search range n+1 steps before. If the allowable range tends to expand quadratically, the intersection of the quadratic curve E, which indicates the boundary condition of the allowable range, and the time n+1 steps before is taken, and the range is defined as the initial search range R. ｎ＋１ Let's say.

[0053] Step 2-2: Initial search range R before n+1 steps ｎ＋１ Conduct route evaluation within the facility.

[0054] An initial search range R including the boundary ｎ＋１ Similarly, by randomly setting search points in a Monte Carlo simulation and performing forward calculations, the initial search range R ｎ＋１ Tolerance range R n steps before ｎ Here too, the initial search range R is determined based on the current natural environment and the results of economic and social policies. ｎ＋１ An upper and / or lower limit may be set for the

[0055] The evaluation function used should be able to evaluate the tolerance range rather than finding the optimal route, so it will suffice to use a function that outputs 1 if the result of the calculation from n+1 steps ago falls within the tolerance range from n steps ago, and 0 if it does not fall within the tolerance range.

[0056] Step 2-3: Initial search range R including the boundary ｎ＋１ Within the range, all search points are within the tolerance range R ｎ , as shown in FIG. 6, the initial search range R ｎ＋１ At the boundary of １ Take the initial search range R ｎ＋１ This expands the future time T ｆ It is possible to search for the allowable range even one step before the final ideal future state.

[0057] Step 2-4: Expanded search range R ｎ＋１ +M １ The same analysis is performed within the search range, and all search points within the search range are within the allowable range R n steps ago. ｎ , as shown in FIG. 6, if it is possible to reach ２ This procedure is repeated for the nth step before the allowable range R ｎ Repeat until the search points that cannot be reached are included in the expanded search range.

[0058] In this case, the expansion range of the margin area can be set arbitrarily. For example, the expansion width can be increased or decreased each time the margin area is expanded. Furthermore, the method of expanding the margin area can be defined for each axis of the state variable. Figure 6 shows a case where the margin area is gradually expanded for a specific state variable.

[0059] Step 2-5: Margin area M １ , M ２ A search range R including ｎ＋１ The results are collected in the n+1 step before the tolerance R ｎ＋１ Let's say.

[0060] In this case, you can either perform mathematical smoothing, or simply define a region using a set of searched values. The former has the potential to draw a relatively clean, continuous boundary, but it inevitably includes points that have not been actually calculated, so there is no guarantee of accuracy near the boundary. The latter is discrete, so the boundary is discontinuous and visibility is difficult, but because it is based on actual calculations, the evaluation results of the tolerance range are reliable.

[0061] So far, the procedure for calculating the allowable range of the state at a given time t has been explained.

[0062] When performing the inverse analysis using the above procedure, the above method may be refined by following the all-at-once method or black box method. Also, instead of scattering points like a Monte Carlo simulation, parameter estimation using a genetic algorithm or evolutionary computation, or boundary condition generation using machine learning may be performed.

[0063] When the tolerance range for the natural environment and socio-economic conditions is calculated using the above procedure, the tolerance range generally widens as one goes back in time. For example, as shown in Figure 7, the tolerance range R is expressed as a cone-shaped area. R ｍ is the tolerance m steps ago, and R ｎ is the tolerance n steps ago.

[0064] 7 is a schematic representation of a pattern in which the tolerance R expands linearly with respect to time, but whether it actually expands linearly or not depends on the state variables. Also, the cone shape is an example.

[0065] Here, a method for limiting the permissible range of a state to a predetermined threshold will be described.

[0066] When analyzing a "tipping point" that would cause the natural environment or human society to irreversibly collapse, the tolerance range can be set to peak at a certain point. In other words, the cone-shaped tolerance range can be set to be bounded by a certain surface.

[0067] For example, it can be set as shown in Figure 8. The cone-shaped tolerance range R is cut at the value Th of the state variable that causes tipping. When the tipping point boundary value Th is exceeded, the natural environment and human society undergo a phase transition, causing changes that cannot be restored on the time scale of human society.

[0068] The tipping point boundary value Th may be defined as a constant value over time, or may be set to vary over time. It may also be defined in such a way that tippings between multiple variables are causally related to each other, resulting in a cascading effect.

[0069] The tipping point boundary value Th is set based on, for example, a known technique (see Reference 2).

[0070] Reference 2: Erwin Lambert and two others, “How northern freshwater input can stabilize thermohaline circulation,” Tellus A: Dynamic Meteorology and Oceanography, November 21, 2016.

[0071] Figure 9 is a phase space diagram showing tipping condition bifurcations for ocean currents, excerpted from Reference 2. If the hatched area is exceeded, an irreversible phase transition occurs, affecting both the natural environment and the economy and society.

[0072] Finally, the allowable range calculation unit 111 passes the allowable range R limited by the tipping point boundary value Th to the allowable range output unit 112 .

[0073] [Example of Acquisition of Allowable Human Activity Paths in the Forecasting Unit 12] In forward scenario exploration, several parameters are set and time evolution based on those parameters is calculated. If all possible parameter sets can be set ideally, the forward calculation results can also be obtained as a distribution.

[0074] When the forward scenario search continues, a future route range H is searched for, as shown in Figure 10. In Figure 10, the future route is expressed as a planar cone-shaped region. R is the allowable range R limited by the tipping point boundary value Th.

[0075] The AND region H' between the allowable range R of the state calculated by the backcasting unit 11 and the future route range H calculated by the forecasting unit 12 ultimately becomes the range of routes that the natural environment and human society can follow (the "allowable human activity route"). Note that Figure 10 shows an example of obtaining an allowable human activity route for one state variable.

[0076] In practice, the parameter sets are set discretely depending on the definition of the narrative, so multiple future routes are output as line graphs within the future route range H, as shown in Figure 11. At this time, the forecasting unit 12 plots the allowable range R at each time point, the time-series future route range H (a line graph representing the future route), and the AND region H' on the same graph and displays them on the screen.

[0077] Since one future path ("human activity path") is calculated for each parameter set, the forecasting unit 12 calculates the future path T ｆ Alternatively, at a predetermined time point T in the future, it is determined whether the value of the future route is within or outside the allowable range R, and the determination result is displayed on the screen. Then, as a result of the determination, a future route that is within the allowable range is output from the search device 1 as an "allowable human activity route."

[0078] [Search Flow for Human Activity Routes] FIG. 11 is a diagram showing a search flow for human activity routes performed by the search device 1. As shown in FIG.

[0079] The backcasting unit 11 calculates, by inverse analysis, the "range of acceptable states" that indicates the range within which the state variables of the natural environment and socio-economy must fall at each future point in time up to the "target ideal future state" (step S1).

[0080] The forecasting unit 12 then receives from the global calculation unit 13 the state of the natural environment and socio-economic situation when calculations are carried out forward from the "current state" using the "evaluation scenario parameters" as a "human activity path," and extracts only the "range of allowable states" from that "human activity path" and outputs it as an "allowable human activity path" (step S2).

[0081] [Notes] (Note 1) The search device 1 includes a backcasting unit 11 that calculates, for each time period, the allowable ranges of both the natural environment and socio-economic conditions that are acceptable to reach an ideal future state by going back in time from the future, and a forecasting unit 12 that calculates the time-series natural environment and socio-economic conditions from the present to the future as a human activity path using parameters related to the natural environment and socio-economy, and outputs the portion of the human activity path that falls within the allowable ranges as an acceptable human activity path.

[0082] (Note 2) The backcasting section calculates the allowable range so as not to exceed tipping points for the natural environment and socio-economic situation.

[0083] (Supplementary Note 3) The forecasting unit outputs the time-series human activity paths and the permissible ranges for each time on the same graph.

[0084] (Supplementary Note 4) When searching for the allowable range n+1 time points before, the backcasting unit expands the search range until it exceeds the allowable range n time points before.

[0085] [Effects] According to this embodiment, the backcasting unit 11 calculates, for each time instant, the allowable ranges of both the natural environment and socio-economic conditions that are acceptable to reach an ideal future state by going back in time from the future, and the forecasting unit 12 calculates the time-series natural environment and socio-economic conditions from the present to the future as a human activity path using parameters related to the natural environment and socio-economy, and outputs the part of the human activity path that falls within the allowable range as an acceptable human activity path, thereby providing a technology that can search for a human activity path that is acceptable for both the natural environment and socio-economy.

[0086] Furthermore, because it has this feature, it is possible to check at each point in time before the target year is reached whether the forward-quantified scenario (the human activity path forward-quantified from the narrative and a parameter set based on the narrative) is within an acceptable range from the perspectives of the natural environment and human society at each point in time in the future.

[0087] According to this embodiment, the backcasting unit 11 calculates the tolerance range so as not to exceed the tipping points of the natural environment and socio-economy, thereby providing a technology that can more appropriately search for human activity routes that are acceptable for both the natural environment and socio-economy.

[0088] According to this embodiment, the forecasting unit 12 outputs the time series of human activity paths and the tolerance range for each time on the same graph, making it possible to monitor whether the natural environment and human society are on a sustainable path and whether humanity is in a state where it can reach a desirable future.

[0089] According to this embodiment, when searching for the allowable range n+1 time points before, the backcasting unit 11 expands the search range until it exceeds the allowable range n time points before. ｆ It is possible to search for the acceptable range even one step before the final ideal future state.

[0090] [Others] The present disclosure is not limited to the above-described embodiments, and various modifications are possible within the scope of the gist of the present disclosure.

[0091] The search device 1 of the present embodiment described above can be realized, for example, by using a general-purpose computer system including a CPU 901, a memory 902, a storage 903, a communication device 904, an input device 905, and an output device 906, as shown in Fig. 13. The memory 902 and the storage 903 are storage devices. In this computer system, the CPU 901 executes a predetermined program loaded onto the memory 902, thereby realizing each function of the search device 1.

[0092] The searching device 1 may be implemented by one computer, or by multiple computers, or may be a virtual machine implemented on a computer.

[0093] The program for the search device 1 can be stored in a computer-readable recording medium such as a HDD, SSD, USB memory, CD, or DVD. The computer-readable recording medium is, for example, a non-transitory recording medium. The program for the search device 1 can also be distributed via a communication network.

[0094] 1 Search device 11 Backcast unit 12 Forecast unit 13 Earth calculation unit 111 Tolerance range calculation unit 112 Tolerance range output unit 121 Future route calculation unit 122 Route confirmation unit 123 Future route output unit 131 Natural environment calculation unit 132 Socioeconomic calculation unit 901 CPU 902 Memory 903 Storage 904 Communication device 905 Input device 906 Output device

Claims

1. A search device comprising: a backcasting unit that calculates, for each time period, the allowable ranges of both the natural environment and socio-economic conditions that are acceptable to reach an ideal future state by going back in time from the future; and a forecasting unit that calculates the time-series natural environment and socio-economic conditions from the present to the future as a human activity path using parameters related to the natural environment and socio-economics, and outputs the portion of the human activity path that falls within the allowable ranges as an allowable human activity path.

2. The search device of claim 1, wherein the backcasting unit calculates the tolerance range so as not to exceed tipping points of the natural environment and socio-economic conditions.

3. The search device according to claim 1, wherein the forecasting unit outputs the time-series human activity routes and the allowable ranges for each time on the same graph.

4. The search device according to claim 1, wherein the backcast unit, when searching the allowable range n+1 time ago, expands the search range until it exceeds the allowable range n time ago.

Images